See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/232536685

Sex Differences in Empathy and

Related Capacities

Article in Psychological Bulletin · July 1983

DOI: 10.1037/0033-2909.94.1.100

CITATIONS READS

759 7,838

2 authors, including:

Nancy Eisenberg
Arizona State University
521 PUBLICATIONS 36,558 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Prosocial Motivation View project

Extensivity Project View project

All content following this page was uploaded by Nancy Eisenberg on 29 May 2014.

The user has requested enhancement of the downloaded file.

Psychological Bulletin Copyright 1983 by the
1983, Vol. 94, No. 1, 100-131 American Psychological Association, Inc.

Sex Differences in Empathy and Related Capacities

Nancy Eisenberg and Randy Lennon
Arizona State University

In this article, the literature on sex differences in empathy (denned as vicarious

affective responding to the emotional state of another) and related capacities
(affective role taking and decoding of nonverbal cues) was reviewed. The literature
was organized and discussed according to method used to assess empathy and
affective role taking. Where appropriate, meta-analyses were also computed. In
general, sex differences in empathy were a function of the methods used to assess
empathy. There was a large sex difference favoring women when the measure of
empathy was self-report scales; moderate differences (favoring females) were
found for reflexive crying and self-report measures in laboratory situations; and
no sex differences were evident when the measure of empathy was either physi-
ological or unobtrusive observations of nonverbal reactions to another's emotional
state. Moreover, few sex differences were found for children's affective role taking
and decoding abilities. Several possible explanations for the pattern of findings
are discussed.

Among the characteristics that people at- the family and society to function are accom-
tribute more frequently to females than to plished. In contrast, women are socialized to
males is the tendency to empathize. This ste- assume an expressive role, that is, to facilitate
reotypic perception has most likely been de- interpersonal harmony within the family
rived from the broader belief that females are unit. Thus, to fulfill their role functions ef-
more nurturant and interpersonally oriented fectively, females, but not males, must be so-
than are males—a stereotype that itself is a cialized to be nurturant, sympathetic, and
natural consequence of traditional feminine empathic.
and masculine roles. Reminiscent of Parsons and Bales' dichot-
Sociological and psychological theorists omy in social roles, two "fundamental mo-
concerned with social behavior generally dalities," agency and communion, were pro-
have not questioned the veracity of sex-role posed by Bakan (1966). Agency is concerned
stereotypes related to empathic reactions. In with the individual, that is, with self-protec-
fact, their conceptualizations are entirely tion and self-assertion. In contrast, commu-
consistent with the notion that females are nion refers to an individual's existence as part
the more empathic sex. For example, soci- of a larger organism, and the individual's
ologists such as Parsons and Bales (1955) sense of being at one with others. Commu-
have attributed differences in males' and fe- nion therefore involves interdependence,
males' behaviors to variations in the tradi- concern for others, and interpersonal consid-
tional roles of the two sexes. According to erations. The concepts of agency and com-
Parsons and Bales, in the family unit men munion seem to reflect masculine and fem-
typically assume an instrumental role; that inine socialization, respectively (Block, 1973),
is, they serve as a liaison between the family and therefore are consistent with the stereo-
and society and see that the tasks needed for type that females are more empathic.
Psychoanalytic theories and conceptual-
The authors wish to thank Warren Eaton for his help- izations are also consistent with the assumed
ful suggestions, Arthur Beaman, Donald Hartmann, and sex difference in empathy. Freud (1925/1961)
William Stock for statistical advice, and the many in- asserted that because females do not fear cas-
dividuals who provided us with unpublished manuscripts tration, they do not resolve the Oedipal com-
and/or additional information and data.
Requests for reprints should be sent to Nancy Eisen- plex as completely as do males. Conse-
berg, Department of Psychology, Arizona State Univer- quently, they develop weaker egos and su-
sity, Tempe, Arizona 85287. peregos. This inferior developmental status
100
SEX DIFFERENCES IN EMPATHY 101

purportedly results in females' relying more to feel the same emotion as another. In con-
on emotion and less on rational principles trast, other investigators define empathy as
than do males. Along somewhat similar lines, sympathy and compassion for another's po-
other psychoanalytic theorists have described sition (a reaction to the other's state that is
females as being more concerned than males not simply emotional matching; e.g., Coke,
with inner space (a concept relating to ma- Batson, & McDavis, 1978), or as a combi-
ternal functioning) and less concerned with nation of emotional matching and sympa-
the external world (Erikson, 1965), as well thetic responding (e.g., Hoffman, 1982;
as being more intuitive (Deutsch, 1944). In- Mehrabian & Epstein, 1972). Researchers
deed, as was noted by Hoffman (1977), there generally have operationalized empathy as an
appear to be no theories that have contra- affective rather than a purely behavioral re-
dicted the stereotype concerning sex differ- sponse to another's affective state and have
ences in empathy. assessed empathy with other-report or self-
report measures, physiological indices, ob-
Definition of Empathy servations of nonverbal indices of reactive-
ness to another's distress, or assessments of
Over the decades, empathy has been de- reflexive crying (crying in reaction to an-
nned in a variety of ways. Some theorists and other's crying).
researchers have viewed empathy as social The primary focus of the present article
insight (Dymond, 1949, 1950), Others have is sex diferences in affective empathy, that is,
defined empathy as the ability to compre- emotional matching and/or sympathetic re-
hend the affective (and sometimes cognitive) sponding. It is this type of empathy that
status of another, that is, as a cognitive ac- seems to reflect most clearly the sex-role ste-
complishment (Borke, 1971, 1973; Deutsch reotype that females are more emotionally
& Madle, 1975; Hogan, 1969). In the past, responsive and sympathetic than males.
the cognitive ability to discern others' inter- Moreover, as was noted by Hoffman (1977),
nal states was sometimes called "sympathy" it is affective empathy that philosophers have
(Cooley, 1902/1956; Mead, 1934) as well as long assumed to play an essential role in so-
empathy, although the term sympathy was cial and moral life (Hume, 1751/1957; Rous-
also used to denote an affective response to seau, 1762/1904; Smith, 1759/1948) and that
another's emotional state (e.g., McDougall, psychologists frequently propose as a media-
1908/1950, 1923; Smith, 1759/1948). More tor of prosocial responding (Aronfreed, 1970;
recently, the ability to discern and compre- Coke et al., 1978; Hoffman, 1976, 1981,
hend others' affective states has been labeled 1982; Isaacs, 1933; Mehrabian & Epstein,
as affective role taking, whereas the ability to 1972; Staub, 1978, 1979; Stern, 1924). Con-
understand another's cognitive status or pro- sequently, because of the theoretical impor-
cessing has been referred to as "cognitive role tance of the concept of empathy, a finding of
taking" (e.g., Ford, 1979; Shantz, 1975; Un- a sex difference in empathy would have rel-
derwood & Moore, 1982). evance not only for the literature on sex dif-
In recent years, many theorists (Feshbach, ferences, but also for the understanding of the
1978; Katz, 1963; Hoffman, 1977; Mehra- broader domain of positive behaviors.
bian & Epstein, 1972; Olden, 1958; Stotland, Theorists have long suggested that affective
1969; Murphy, 1937) have defined empathy role taking can provide the basis for empathic
in more affective terms, that is, as affective responding, especially situationally appro-
responsiveness to another's emotional state priate empathic responding (Feshbach, 1978;
(frequently as the result of cognitively iden- Hoffman, 1976). Although empathic arousal
tifying the other's state). This responsiveness may occur without affective role taking (Hoff-
can be described in two different ways, al- man, 1982; Note 1), the ability to identify the
though most theorists do not make this dis- emotions of others, especially when those
tinction (cf. Wispe, 1968). Some investigators emotions differ from one's own affective
define empathy as vicarious matching of state, is viewed as an important component
emotional responses (Feshbach & Roe, 1968; of most empathic responding (Feshbach,
Stotland, 1969). Thus, to empathize means 1978). Thus, although the primary purpose
102 NANCY EISENBERG AND RANDY LENNON

of the present article was to review the lit- sensitivity. He reviewed nine articles (includ-
erature on sex differences in empathy, a sec- ing 16 samples), all but one involving chil-
ondary purpose was to examine the research dren or infants. Females tended to score
on sex differences in affective role taking. higher on empathy in all of the studies; the
differences were significant in 3 of 16 samples
Prior Reviews on Sex Differences and marginally significant for 3 other sam-
in Empathy ples. On the basis on these findings, Hoffman
concluded that females are indeed more em-
Sex differences in empathic responding pathic than males.
and social sensitivity were first systematically At the time that Hoffman wrote his review,
examined by Maccoby and Jacklin (1974). there were relatively few studies on empathy.
They reviewed 29 papers (including 31 sam- Since then, considerably more research has
ples) related to "empathy," and concluded been conducted. Moreover, most of the stud-
that there was no evidence of a sex difference. ies reviewed by Hoffman were conducted
In most of the studies examined, there were with one of two methods. In the past few
no significant differences in the responses of years, researchers have assessed empathy
the two sexes; the sex differences that were with a variety of methods including story/
found in the remaining studies were divided picture techniques, physiological measures,
nearly equally in favor of males and females. facial, gestural, or vocal indices, and self- or
However, included among the references re- other-report measures. Because each method
viewed were studies pertaining not only to has its own set of shortcomings, and different
emotional empathy, but also to a variety of methods may not be equally valid, in the
other behaviors related to social sensitivity, present review the research on sex differences
for example, affective role taking, accuracy in empathy was examined separately for each
of guessing how others rate themselves, and mode of assessment.
ability to take different dramatic roles. In
only four of the references reviewed by Mac- Methods of Analysis
coby and Jacklin was empathy defined as a
predominantly affective response (Craig & Description of the Data Set
Lowery, 1969; Feshbach & Feshbach, 1969,
2 samples; Feshbach & Roe, 1968; Simner, To provide a comprehensive review of
1971, 2 samples); one other study was not studies related to sex differences in empathy,
described as pertaining to affective empathy all relevant empirical research that could be
although empathy was examined in that located was included in the review. This in-
study (Hamilton, 1973). cluded published studies, articles being sub-
In two more recent reviews, the conclu- mitted for publication, and those disserta-
sions of Maccoby and Jacklin regarding sex tions for which adequate detail could be ob-
differences in empathy were reexamined. In tained.
her 1976 review, Block (1976) did not define To aid in examination and interpretation
empathy (or sensitivity to social cues) differ- of the research, the empirical data were or-
ently from Maccoby and Jacklin, and she ganized according to the method by which
found that sex differences in 23% of the stud- empathy was assessed. This organization re-
ies on "empathy" favored females while 10% sulted in seven groupings of research, each
favored males. On the basis of these figures, involving one of the following seven measures
Block did not challenge the conclusions of of empathy: (a) infants' reflexive crying in
Maccoby and Jacklin regarding sex differ- response to another's distress; (b) individuals'
ences in empathy. reports to the experimenter of emotional re-
In contrast to Block and Maccoby and sponsiveness after hearing stories or viewing
Jacklin, Hoffman (1977), in his review of sex pictures containing information regarding a
differences in empathy, differentiated be- hypothetical other's affective state; (c) self-re-
tween measures of empathy (defined as emo- port (via pencil and paper measures) of
tional responsiveness to another's affective emotional responsiveness in simulated dis-
state) and indices of role taking and social tress situations; (d) observers' ratings of in-
SEX DIFFERENCES IN EMPATHY 103

dividuals' facial, gestural, and/or vocal re- Methods of Evaluation

actions to another's emotional state; (e) sub-
jects' physiological responsiveness to another's There are a variety of ways in which one
predicament; (f) individuals' responses on can review a pattern of empirical findings.
self-report scales specifically designed to Our approach will be both qualitative and
measure empathy; or (g) report by others of quantitative. Thus, the results of the various
individuals' empathy. types of studies are discussed, as well as the
In reality, the categories of research de- strengths and weaknesses of each method.
scribed above differ somewhat from one an- Where appropriate, the results related to a
other with regard to definition of empathy, specific methodology will be summarized
as well as with regard to method used to as- with a meta-analysis.
sess empathy. In some studies, empathy was The term meta-analysis has been used to
defined as the matching of one's own emo- refer to any statistical technique used to com-
tional responses to that of another (most bine the results of independent studies (Glass,
studies in Group b); in others, empathy was 1976; Glass, McGaw, & Smith, 1981). It is
defined, at least in part, as sympathy or con- a method of integrating the results of a num-
cern for others (most studies in Groups c, f, ber of different studies that are believed to
and g). Of course, when an individual's contain data related to a single hypothesis.
emotional responsiveness is operationalized When the results of research are inconsistent,
as physiological or facial/gestural/vocal re- a meta-analysis can be used to evaluate the
sponses to another's emotional state (Groups overall outcome of the data from the com-
d and e), it is difficult to determine whether bined studies.
an affective response reflects emotional A relatively simple type of meta-analytic
matching or sympathy. Similarly, in those procedure is the Stouffer method (Mosteller
cases in which empathy was operationalized & Bush, 1954; Rosenthal, 1978; Stouffer,
as an individual's self-report of a global neg- Suchman, DeVinney, Star, & Williams, 1949).
ative state (e.g., sadness), it is impossible to This method has been used and discussed in
determine if the negative reaction reflected recent reviews related to sex differences in
emotional matching or sympathy. conformity and the association between al-
As will become evident in the review, most truism and perspective-taking (Cooper, 1979;
of the various methods for assessing empathy Underwood & Moore, 1982).
have been used with subjects spanning a nar- In brief, the Stouffer method is based on
row age range. Studies of reflexive crying in the fact that the sum of independent standard
response to another's crying have been con- normal deviates (Z scores) is another stan-
ducted only with newborns. Techniques in dard normal deviate with its variance equal
which empathy is operationalized as individ- to the variance of all the standard normal
uals' reports to the experimenter of emo- deviates that were summed. Thus, a new
tional responsiveness in reaction to a hypo- standard normal deviate, which is formed
thetical story have been used primarily with from summing a number of standard normal
children. In contrast, self-report of respon- deviates, can be computed with the following
siveness to a staged or videotaped situation formula:
via pencil and paper measures (rather than
direct report to the experimenter) has been i
used most frequently with adults, Physiolog- '

ical instruments have been used primarily where Zma is the new standard normal de-
with adult samples, while facial/gestural/vo- viate resulting from the meta-analysis; Z sl ,
cal measures have been used solely with chil- . . . , Zsn are the standard normal deviates for
dren. Finally, researchers have used tests or each included study; and Nsi , . . . , Nsn equals
scales of sympathetic responsiveness to assess the number of subjects in each included
the personality trait of empathy (rather than study. The level of significance of Zma can be
empathy as a response to a specific situation) obtained simply by consulting the table of
with adults and children as young as ap- standard normal probabilities in any book
proximately age 6. of statistics.
104 NANCY EISENBERG AND RANDY LENNON

The formula presented above is only one inclusive, it was used in the meta-analysis.
of two variations on the Stouffer method (see For instance, if a subscale of a measure of
Cooper, 1979). With this method, the num- empathy as well as a total composite score
ber of subjects in each study is used to weight (comprising the results from all the subscales)
the contribution of each study to the formula was included in an article, all measures were
so that studies involving larger Ns are weighted included in our summary tables, but only the
more heavily than studies with fewer subjects. composite score was used in any meta-anal-
Because the number of subjects varies con- ysis. If no one measure of sex differences was
siderably across studies of sex differences in considered superior or superordinate, the
empathy, we chose to use the formula pre- standard normal deviates for the various
sented above rather than the alternative, measures were averaged, and the mean was
which does not weight by N, used in the meta-analysis.
To compute the formula, all that is needed When one is attempting to review research
is the one-tailed probability value for the rel- on an issue such as sex differences with regard
evant finding in each study. Thus, for two- to a specific behavior, there is another prob-
tailed tests (including all Fs from analyses of lem that is likely to complicate matters. Fre-
variance), the p must be converted to a one- quently, data related to sex differences are
tailed probability (cf. Mosteller & Bush, embedded in studies in which the issue of sex
1954; Rosenthal, 1980). This probability differences was of minor importance. Thus,
value can then be converted to a standard if no sex difference was obtained in a partic-
normal deviate. Of course, standardized ta- ular study, this finding may have been re-
bles can be used to convert F or t statistics ported in insufficient detail or not at all. If
to standard normal deviates if the degrees of a researcher merely reports that "there was
freedom and/or number of subjects for a no significant sex difference in empathy," but
study are known. Similarly, when data are does not supply any further information, it
reported in the form of correlation coeffi- is impossible to convert the data to a standard
cients, a standard normal deviate can be ob- normal deviate. The best that one can do in
tained with the Zr to Z transformation. Fi- such circumstances is to assume that findings
nally, Mann-Whitney Us can be converted of no difference average out, across studies,
to standard normal deviates by means of two to a standard normal deviate of 0, with a
formulas presented by Ferguson (1966, pp. probability level of .5 (cf. Cooper, 1979). In
359-360). All of these techniques were used the present review, we adopted this assump-
in the present review. tion. However, when the bulk of data related
Frequently there are several measures of to sex differences in empathy was merely re-
sex differences in empathy in a single study. ported as nonsignificant, we chose not to
Because of the assumption of independence compute a meta-analysis at all (rather than
of the measures involved in a single meta- to compute it using many inexact estimates
analysis, only one statistic from any given of standard normal deviates).
sample (not article or study) can be included As Rosenthal (1979) pointed out, it is
in a specific meta-analysis. Thus, in any highly unlikely that any literature review will
meta-analysis, one has to devise methods of uncover all relevant studies related to the
dealing with multiple measures. In the pres- hypothesis being tested. The problem (called
ent review, when multiple measures of em- the "file-drawer" problem by Rosenthal) is
pathy were obtained with a single sample that when researchers fail to obtain expected
using different methods (e.g., physiological results, these data are usually not published.
measures and self-report), the results of the If the unpublished studies tend to be ones in
various analyses were reviewed separately as which the null hypothesis was not rejected,
each specific mode of assessing empathy was the literature that is available to a reviewer
being considered. In those cases in which a is likely to overestimate the strength of the
researcher reported two or more analyses in- relationship in question.
volving a single method, one of two methods Although there is no ready solution for this
of handling the data was used. First, if one problem, Cooper (1979, p. 147) suggested a
measure was deemed more adequate and/or method for estimating the magnitude of the
SEX DIFFERENCES IN EMPATHY 105

file-drawer problem. Cooper provided a for- empathic response. Because data on reflexive
mula for calculating the "fail-safe number," crying show a more consistent pattern of sex
or the number of studies with null results that differences than most other types of data,
would be necessary to decrease the meta- they deserve careful scrutiny.
analysis probability below significance (p < We could locate only three published pa-
.05). This statistic should be used only when pers (including data on seven different sam-
a significant overall probability is reported, ples) related to sex differences in reflexive
and we have done so in those cases. crying. Simner (1971) conducted four of
Finally, as was suggested by Cooper (1979) these studies. He found that infants cried
and Glass et al. (1981), a measure of the more in response to another newborn's cry
strength of the relations (between sex and than in response to either white noise, a 5'/z-
empathy) was included. Such a measure has month-old's cry, or a computer simulated
been called an indicator of "effect size." Be- cry. Moreover, he found that females exhib-
cause the usual criterion for determining the ited somewhat more reflexive crying in all
significance of a research finding is a function studies, with the difference being significant
of sample size (O'Brien & Shapiro, 1968), it in only one of the four studies and nonsig-
is desirable to present a measure of the mag- nificant in all the others (see Table 1). Sex
nitude of the effect that is free of sample size. differences in crying in the various other ex-
There are a variety of measures that can perimental conditions (e.g., in response to
be used to compute effect size (Cohen, 1977). white noise), or in the no noise control were
In the present review, the measure chosen was generally not reported or discussed.
that of Glass et al. (1981, see pp. 126-131 In a more recent study involving infants
for exact formulas for computing this index). somewhat younger than in Simner's research,
According to Cohen, an effect size of .20 is Sagi and Hoffman (1976), like Simner, found
considered small, .50 is moderate, and .80 is that females tended to exhibit more reflexive
large. crying than did males. They found no sex
When more than one finding was reported differences in crying in response to either si-
for a single sample, the procedures described lence or a synthetic cry. On the basis of these
above for the computation of Stouffer's Z data, Hoffman (1977) suggested that girls ex-
score were used to compute a measure of hibit more empathic tendencies than do boys
effect size. Moreover, in accordance with as early as the first hours of life.
Cooper (1979), we assigned an effect size of In the most recent research on reflexive
0 to studies for which no standard normal crying, Martin and Clark (1982) examined
deviate could be computed, and a finding of reflexive crying in response to cries of a new-
"no difference" was reported. born male, an 11-month-old male, a 16-day-
To summarize, in this review the literature old male chimpanzee, or the child's own
will be examined qualitatively, and, where tape-recorded cries. They found no sex dif-
possible, with a quantitative meta-analysis ferences in reflexive crying in two separate
package. As part of the quantitative review, studies. Children cried more in response to
we will report the following measures: (a) the a newborn male's cry than in reaction to their
Stouffer's Z, (b) a fail-safe number, and (c) own cries or those of an 11-month-old child
a measure of effect size. or a chimpanzee.
According to a meta-analysis performed
Review of the Empirical Literature for the results of the seven experiments con-
Research on Infants' Reactions to Another's ducted by Simner, Martin and Clark, and
Crying (Reflexive Crying) Sagi and Hoffman, the Z score was signifi-
cant, Z = 2.12, 1-tailed/J < .017. The mean
The research on reflexive crying among effect size was .27,' and the fail-safe number
infants was cited by both Maccoby and Jack-
lin (1974) and Hoffman (1977) as pertaining
' Effect sizes for each study were recomputed with
to the issue of sex differences in empathy. The Hedges' (1982) formulas for unbiased estimators of effect
assumption has been that infants' crying in size (Formulas 4 and 6) and average effect size was re-
response to another's cry reflects a primitive computed with Hedges' new formula designed to weight
106 NANCY EISENBERG AND RANDY LENNON

Table 1
Sex Differences in Reflexive Crying
Sex with Effect
Reference Age &N Measure One-tailed p advantage size

Martin & Clark

(1982)
Experiment 1 M = 18 hrs. Vocalizations in response to ns F (0)
(40) tape of crying infant (self
or newborn male)
Experiment 2 M = 29 hrs. Crying in response to tape of ns NR (0)
(30)a newborn or 11 -month-old
male
Sagi & Hoffman M = 34 hrs. Crying in response to tape of .06" F .72
(1976) (20) crying female infant
Simner
c-
(1971)
Experiment 1 64-76 hrs. Crying in response to tape of .039 F .37
(94) crying female infant
Experiment 2 67-77 hrs. .121° F .48
(25)
Experiment 4 M = 59 hrs. .350 F .15
(30)
Ancillary M = 64-74 hrs. .196 F .17
data (100)

Note. F = female; NR = not reported.

a
It is not entirely clear whether 20 or 30 infants were included in the analysis of sex differences (i.e., if all conditions
were included in this analysis).
b
According to a personal communication, this was a one-tailed test.
c
This was computed by assuming there were approximately equal numbers of male and female subjects (a reasonable
assumption given the information provided).

was approximately 5. Thus, despite some in- Clark, 1982; Simner, 1971), one reason that
consistencies in the empirical^ data, females females cried more than males in the Simner
do seem to exhibit somewhat more reflexive and Sagi and Hoffman studies may have been
crying than do males. because the stimulus cry was more similar
Although this difference in reflexive crying to their own cry than to that of the male
may, indeed, reflect a sex difference in em- subjects. In the two studies in which the stim-
pathic responding, such a conclusion may be ulus cry was that of a male, there were no sex
premature for several reasons. First, there is differences in reflexive crying. Because males
an important methodological limitation in did not cry more than females in reaction to
the research. In all of the studies described a male infant's cry, the factor of sex of the
above except those of Martin and Clark, the stimulus infant cannot entirely account for
stimulus cry was that of a female infant. Be- the pattern of findings. Nonetheless, the use
cause newborns appear more likely to exhibit of female stimulus infants in most research
reflexive crying in response to a cry similar may have accentuated any real sex difference.
(but not identical) to their own (Martin & Even if there were no methodological prob-
lems with the research on reflexive crying,
there would still be obstacles to interpreting
each study when computing mean effect size (Formula
11). Mean effect size for the studies involved in the re- the data. It certainly is not clear that reflexive
flexive crying, picture/story, and self-report question- crying implies empathic responding. It is pos-
naire meta-analyses (Tables 1, 2, and 6), when computed sible that female infants, because of sex dif-
with Hedges' formulas, were .34, .11, and .91, respec- ferences in certain sensory capacities or
tively. When doing these computations, if the number
of male versus female subjects was not available, it was
thresholds (Friedman & Jacobs, 1981; Lipsitt
assumed that there were equal numbers of males and & Levy, 1959; Singer, Westphal, & Niswan-
females. der, 1968), are more responsive than are
SEX DIFFERENCES IN EMPATHY 107

males to specific auditory features of the hu- "How do you feel?", "Tell me how you feel?",
man cry. Furthermore, since Simner (1971) or "How did that story make you feel?"
found little stability in reflexive crying over (Feshbach & Roe, 1968; Roe, Note 2).
a 24-hour period (although somewhat more Verbal responses to the PASTE or deriv-
stability was found for females than for ative measures have typically been scored in
males), it does not seem likely that sex dif- one of three ways. Feshbach and Roe devised
ferences in reflexive crying are precursors of two of the three systems of coding; Levine
stable differences in empathic reactions at a and Hoffman (1975), the third. With the
later age. Feshbach and Roe exact scoring system
In summary, although female newborns (henceforth called ES), responses that specif-
seem to exhibit more reflexive crying when ically match the affect portrayed in a specific
exposed to the cry of another newborn infant, story/slide sequence (e.g., for a sad narrative:
the significance of this sex difference is un- "I feel sad") are assigned a score of 1, and
clear. More controlled research (using chil- all other responses receive a score of 0. Ac-
dren of both sexes as stimuli) and longitu- cording to the second system devised by Fesh-
dinal research is needed to clarify the issue. bach and Roe, the general system (henceforth
called GS), responses that specifically or gen-
Research Involving Picture/Story erally match the affect portrayed in the story
Techniques and Interview (e.g., for sadness: "I feel sad" or "I feel bad")
By the Experimenter are assigned a score of 1, and all other re-
sponses are scored 0. Finally, according to
The most popular method used to study Levine and Hoffman's system (henceforth
empathy in children has been what can be called the "3-point system"), a specific match
called the picture/story technique. In es- between the child's and the story character's
sence, with this method the individual is ex- affect is scored 2; a general match is scored
posed to another's affective state via pictures 1; and incorrect responses are scored 0.
and/or a story concerning a hypothetical Other researchers have modified the PASTE
story protagonist; that is, via relatively non- procedure and/or scoring system (e.g., Eisen-
lifelike stimuli (with the exception of Fesh- berg-Berg & Lennon, 1980; Hughes, Tingle,
bach, 1982). Then the child is requested to & Sawin, 1981; lannotti, Note 3; Miller, Note
indicate his or her own emotional state (usu- 4). For instance, sometimes only a subset of
ally by verbal report to the experimenter or the four affects (e.g., happiness and sadness)
by pointing to pictures of facial expressions). has been used (Eisenberg-Berg & Lennon,
In most of these studies, a match between the 1980; Hughes et al., 1981). In other research,
child's affect and that of the story protagonist the scoring procedures have been modified,
is viewed as an empathic response. (for example, children have indicated their
Most of the research of this type has been affective state with nonverbal, pointing re-
modeled upon the work of Norma Feshbach sponses (e.g., Eisenberg-Berg & Lennon, 1980;
and her colleagues. In one of the relatively lannotti, 1978, Note 3)), or different types of
early studies on empathy in children, Fesh- empathy have been scored (lannotti, 1978,
bach and Roe (1968) developed an instru- Note 3; Staub & Feinberg, Note 5). Further-
ment called the PASTE (Feshbach and Roe more, in recent research, Feshbach and her
Affective Situations Test for Empathy) to as- colleagues (1982) have used videotape se-
sess preschoolers' and young elementary quences rather than slides (see Table 2).
school children's empathy. The instrument There has been considerable discussion
consists of eight short stories, each accom- concerning the psychometric characteristics
panied by three slides depicting the story of picture/story measures of empathy (e.g.,
events. In each narrative, the story protago- Eisenberg-Berg & Lennon, 1980; Hoffman,
nist (a child) is involved in an emotion-pro- 1982; Sawin, Note 11). Some researchers
voking event. There are two narratives de- have suggested that the demand characteris-
picting each of the following four emotions: tics inherent in the PASTE procedure, caused
happiness, sadness, fear, and anger. After by an adult repeatedly asking the child how
viewing any given story, the child is asked he or she feels, undermine its validity. It is
108 NANCY EISENBERG AND RANDY LENNON

Table 2
Sex Differences in Empathy as Measured With Picture/Story Techniques (Such as the PASTE)
Sex of One-tailed Sex with Effect
Reference Age & W Measure experimenter P advantage size

Barnett, King, P&K PASTE M ns F (0)

Howard, & (54) 3 pt.
Dino
(1980)
Barnett, G 1 PASTE M ns M (0)
Mathews, & (84) 3 pt.
Howard
(1979)
Brehm, Powell G1 Self-report of reactions F
&Coke (67) to sad audio tape
(in press) ES .050 F 4fi
,*ru
GS .019 F
Bryant G 1 PASTE F &M
(1982) (56) ES ,241s F
GS .234" F .19
Fay 6-8 yrs. PASTE F ns NR (0)
(1971) (60)
Eisenberg-Berg P Modified M
& Lennon (51) PASTE
(1980) Verbal .054 M
Report -.57
Nonverbal .011 M
Response
Eisenberg-Berg & P PASTE M .050 M -.60
Lennon (32) 3pt.
(Note 6)
Feshbach G 3 &4 Audio-visual test of NR
(1982) («100) empathy
Self-report for:
Dysphoric emotions ns NR /A\
(V)
Euphoric emotions NR F
Feshbach & 4-5 yrs. PASTE
Feshbach (48) ES Fb .03 F .55
(1969) 6-7 yrs. ES Fb .21 F .2T
(40)
Feshbach & Roe 6-7 yrs. PASTE F
(1968) (46) ES (same sex stimuli) .025 F
ES (opposite sex ns M .15
stimuli)
GSd ns M
Hughes, Tingle, K & 7-8 yrs. Modified NR ns NR (0)
& Sawin (24 each) PASTE
(1981)
lannoti P Pictures & stories F&M
(Note 3) (52)
Affect matched to ns M
protagnoist's affect (0)
Affect matched to ns F
situational story
cues
Lennon, P PASTE F&M
Eisenberg, & (80)
Carroll
(Note 7)
ES .199 F
GS .198 M
3-pt. .444 F
SEX DIFFERENCES IN EMPATHY 109

Table 2 (continued)

Sex of One-tailed Sex with Effect

Reference Age& N Measure experimenter p advantage size
Lennon, P PASTE F&M .103 M -.44
Eisenberg, & (35) 3pt.
Carroll
(Note 8)
Levine & P PASTE Fa ,06£ F .42
Hoffman (80) 3-pt.
(1975)
Marcus P PASTE F ,025a F .76
(Note 9) (32) 3-pt.
Marcus, Telleen, P PASTE NR
& Roke (32) ES ns NR
(1979) GS ns NR (0)
Miller 9-10 yrs. Modified F .170 F .24
(Note 4) (66) PASTE
Roe 6-7 yrs PASTE F .025e F .50
(1977) (64)
Greek
childrenr
Sawin, 6-7 & 8-9 yrs. PASTE F&M
Underwood, (128) ES ns NR (0)
Weaver, & 3pt. ns NR
Mostyn
(Note 10)
Staub & Feinberg G3&4 Photographs of people NR
(Note 5) (51)a in different
situations
Parallel empathy11 .189a F .26
Reactive empathy11 Significant"1' F

Note. For Age & N, P = preschool (usually 4-5 yrs.); K = kindergarten; G = grade. For Measure, PASTE = Feshbach
& Roe (1968) Affective Situations Test; 3-pt. = Levine & Hoffman (1975) 3-point coding system; ES = Feshbach
& Roe (1968) exact scoring system; GS = Feshbach & Roe (1968) general scoring system. F = female; M = male;
NR = not reported.
a
Personal communication from author.
b
Personal report from Kiki Roe.
c
Computed from Ms and SDs given in the article.
"Reported in Roe (1977).
e
According to personal communication from the author, this is a one-tailed test. These apparently are the same
data reported in Hoffman & Levine (1976).
r
The American children in this story were those in Feshbach & Roe (1968).
8
The F for the simple main effect was not reported, although the gender difference for Greek children was reported
to be significant. Thus, the F was estimated to be significant at p < .05 (p < .025 for one-tailed test).
h
Parallel empathy = reports of feelings similar to those of story protagonist; reactive empathy = report of sympathetic
feelings.
' An exact F could not be computed from the reported data. However, since parallel empathy is more analogous
to the measures in other studies, the data for parallel empathy were used in the meta-analysis.

also possible that the short hypothetical sto- of subject with sex of experimenter. As is ev-
ries that comprise these instruments are un- ident in Table 2, in all studies in which fe-
likely to elicit empathy in young children. males scored significantly or nearly signifi-
Moreover, there appears to be a methodolog- cantly higher than did males, the experi-
ical problem with the administration proce- menter was female (Brehm, Powell, & Coke,
dures used in many studies involving the in press; Feshbach & Feshbach, 1969; Fesh-
PASTE and derivative measures. Specifically, bach & Roe, 1968; Levine & Hoffman, 1975;
the assessment of sex differences in empathy Roe, 1977). In contrast, in all studies in
may be confounded by the interaction of sex which males scored higher than females (Ei-
110 NANCY EISENBERG AND RANDY LENNON

senberg-Berg & Lennon, 1980, Note 6), the rho = .80, p < .001; subjects scored higher
experimenter was male, Moreover, in a recent when interviewed by same-sex experi-
study with 80 preschoolers in which sex of menters.
subject and sex of experimenter were system-
atically varied, there was evidence of a Sex Self-Report Measures of Empathy in
of Subject X Sex of Experimenter interaction Simulated Emotional Situations
for performance on the PASTE (Lennon,
Eisenberg, & Carroll, Note 7), especially The set of studies involving self-report of
when only the emotions understood by chil- empathic reactions to simulated situations
dren were considered. Children interviewed differs from studies involving picture/story
by same-sex experimenters scored higher techniques in the following ways: (a) the
than subjects interviewed by opposite-sex ex- emotion-provoking situations were presented
perimenters. Apparently, children attend with videotapes (e.g., of a crying infant or a
more to same-sex than opposite-sex experi- person in need of help) or realistic situations
menters or fear same-sex experimenters less. (e.g., a simulated jury trial, a real person
Alternatively, children may be more moti- being administered electric shock); (b) sub-
vated to seek approval (and thus, to provide jects were frequently asked to report sym-
"correct" answers) from same-sex experi- pathetic reactions (feelings of concern) as well
menters. Whatever the reason for the inter- as (or in place of) affective reactions that ex-
action, the findings of Lennon et al. render actly matched the story protagonist's affect
interpretation of data related to sex differ- (Archer, Foushee, Davis, & Aderman, 1979;
ences in empathy as assessed by the PASTE Batson, O'Quin, Flutz, Vanderplus, & Isen,
and similar methods hazardous. Note 12; Prodi & Lamb, 1978; Prodi, Lamb,
Having outlined the potential limitations Leavitt, & Donovan, 1978; Prodi, Lamb,
of research conducted with picture/story Leavitt, Donovan, Neff, & Sherry, 1978;
techniques, we will next review the results of Zahn-Waxier, Friedman, & Cummings, Note
this research. According to a meta-analysis, 13); and (c) the subjects reported their aff-
females score significantly higher on picture/ fective reactions by means of pencil and pa-
story measures of empathy than do males, per measures, not via verbal report directly
Z = 1.79, 1 -tailed p < .04. However, the dif- to the experimenter (except in the study of
ference is small, and the fail-safe number is Zahn-Waxier et al., Note 13).
less than 4. Moreover, because effect size, Eleven articles containing 14 studies of the
unlike the Stouffer Z, is not a function of type described above were located (see Table
sample size (and the Ns tended to be rela- 3). In these articles, there were 2 samples of
tively large for studies favoring females) the children and 12 of adults. In most of these
mean effect size across studies was a mea- studies, the authors reported that there were
ger +.10 (see Footnote 1), implying a small no sex differences and did not provide spe-
female advantage. cific data.
Because of the apparent interaction of sex With regard to the two studies involving
of experimenter with sex of subject in influ- children (both of which concerned respon-
encing empathy scores, the finding of a sex siveness to a distressed baby), no sex differ-
difference in studies of empathy involving ence was reported in one study (Prodi &
picture/story techniques must be qualified by Lamb, 1978), whereas girls reported more
an awareness that most of these studies were sympathetic reactions in the second (Zahn-
conducted by female experimenters. In order Waxler et al., Note 13). In the studies with
to calculate the degree of association between adults, there were no sex differences in seven
sex of experimenter and sex differences in studies; males reported significantly more
empathy scores, sex of experimenter (coded empathy on one of two measures in one
1 = male, 2 = male and female, 3 = female) study; and females reported more empathy
was correlated with effect size (for the 18 in four studies (although in two of these, sig-
samples for which we could ascertain sex of nificant differences were found for only a sub-
experimenter). The resulting correlation sample of dependent variables). Moreover, it
coefficient was highly significant, Spearman should be noted that in two of the studies in
SEX DIFFERENCES IN EMPATHY 111

which females reported more empathic re- or anxiety), one would also predict that fe-
actions, their facial or gestural (Zahn-Waxier males would exhibit greater physiological re-
et al., Note 13), and/or physiological reac- activeness. However, there is little evidence
tions (Craig & Lowery, 1969) were not con- of a sex difference in physiological response
sistent with self-reported affect. to another's emotional distress. The only
In many of the studies listed in Table 3, study in which there was a significant sex
the researchers reported insufficient data for difference was Craig and Lowery (1969), and
transforming the results to a standard normal in that study, males exhibited more galvanic
deviate. Thus, a meta-analysis was not com- skin response (but not deceleration of heart
puted for these studies. Consequently, given rate) than did females.
the limited data available, the most that one In the Craig and Lowery study, there were
could conclude is that there is a tendency for four experimental conditions, two in which
females to report more empathic reactions a confederate was ostensibly shocked and two
than do males in reaction to others' simulated control groups. Across conditions, the main
affective states. However, the self-report mea- effect of sex was significant but there was no
sures frequently are not consistent with other Sex X Condition interaction. Males appar-
measures of empathic responding. Thus, the ently exhibited greater physiological respond-
results of the data regarding sex differences ing even when the confederate was not being
in self-report of empathy in simulated situ- shocked. Thus, it is likely that the greater
ations are far from conclusive. physiological responding among males was
due to factors other than, or in addition to,
Physiological Measures of empathizing.
Empathic Responding Because most of the researchers who con-
ducted studies involving physiological indices
Few researchers have used physiological of empathy did not report specific statistics
measures to examine sex differences in em- related to sex differences, it would have been
pathy. Only seven research reports (including inappropriate to compute a meta-analysis.
nine samples) were located (see Table 4). The Nonetheless, it is highly unlikely that such
circumstances in which physiological respon- an analysis would have been statistically sig-
ses have been examined are situations in- nificant. It should be noted, however, that at
volving (a) a crying infant (Frodi & Lamb, this time it is difficult to interpret even null
1978; Frodi, Lamb, Leavitt, & Donovan, results in research involving physiological
1978; Frodi et al., 1978; Murray, 1978); (b) measures. This is because empathic respon-
person being shocked (Berger, 1962, two stud- ses might be evident in some physiological
ies; Craig & Lowery, 1969), and (c) person responses but not others, and cognitive pro-
taking a test (Stotland, 1969, two studies). In cesses might distort or mask physiological
all except one study (Frodi & Lamb, 1978), reactions. Moreover, due to individual dif-
the subjects were adults. ferences, it is unclear how to measure phys-
When a person responds physiologically to iological reactions to specific stimuli (see
another's distress, it is difficult to determine Lang, 1971, for discussion of these issues).
what the physiological reaction represents.
Physiological responsiveness may reflect em- Facial, Vocal, and Gestural Measures
pathy, but it might also represent fear, irri- of Empathy
tation, or personal distress due to the aver-
siveness of a situation. Nonetheless, if fe- As for the data concerning physiological
males were significantly more empathic than measures of empathy, the research related to
males, one would predict a sex difference in sex differences in facial, gestural, and vocal
subjects' reactions to viewing another's dis- indices of empathy is relatively limited. Only
tress. Moreover, if females were to experience six articles, all involving children, could be
more fear or nonempathic distress in situa- located (see Table 5).
tions in which another person was under The studies differ considerably in method.
stress (as might be expected from stereotypes In two of the studies, children's vocal and
concerning sex differences with regard to fear facial cues were rated while they were verbally
112 NANCY EISENBERG AND RANDY LENNON

responding to the PASTE (Feshbach and Roe fant in a seemingly natural situation were
Affective Situations Test; Marcus, Note 9; observed (Zahn-Waxier et al., Note 13).
Sawin, Underwood, Weaver, & Mostyn, Note The only studies in which there were sex
10). In three other studies, researchers rated differences were those in which children were
children's facial and/or gestural reactions observed while responding to the PASTE.
while they observed films of others in happy Both Sawin et al. (Note 10) and Marcus
and/or sad situations (and were not being (Note 9) found that girls exhibited more vo-
questioned regarding their affective state; cal responsiveness than did boys; sex differ-
Hamilton, 1973; Leiman, Note 14; Lennon, ences in facial responsiveness were found by
Eisenberg, & Carroll, Note 8). In the one re- Sawin et al. but not Marcus. Since there were
maining study, children's facial and gestural no consistent sex differences in facial, ges-
reactions to a tape recording of a crying in- tural, or vocal responsiveness in any of the

Table 3
Sex Differences in Self-Report of Empathic and Sympathetic Reactions in Simulated Situations
One-tailed Sex with
Reference Age &N Measure P advantage

Archer, Foushee, Davis & College Ratings of mood (sadness) ns NR

Aderman (1979) (121) in response to
simulated jury trial
Report of role taking .030 M
Batson, O'Quin, Fultz, Vanderplus,
&Isen (Note 12)
Experiment 1 College Reactions to seeing
(40) another shocked
(composite scales)
Empathy scale" ns NR
Empathy scale minus ns NR
distress scale
Experiment 2 College Empathy scale" ns NR
(40) Empathy scale minus ns NR
distress scale
Experiment 3 College Empathy scale" .025 F
(40) Empathy scale minus ns NR
distress scale
Craig & Lowery (1969) College Ratings of reaction to .004 F
(28 in this analysis) seeing another shocked
Frodi & Lamb (1978) 8-14yrs. Responses on Mood ns NR
(32) Adjective Checklist
after exposure to
videotape of crying
infant b
Frodi, Lamb, Leavitt, & Donovan Parents Responses on Mood ns NR
(1978) (96) Adjective Checklist
after exposure to
videotape of crying
infant'
Frodi, Lamb, Leavitt, Donovan, Parents Responses on Mood ns Md
Neflf, & Sherry (1978) (64) Adjective Checklist
after exposure to
videotape of crying
newborn (full term or
premature)0
Murray (1978, in press) College Ratings of sympathy in <.001 F
(139) reaction to audiotaped
cries of newborns
Stotland (1969; Reanalysis of College Ratings of tension while NR Fc
Stotland & Dunn, 1963) (125) observing another
taking a test
SEX DIFFERENCES IN EMPATHY 113

Table 3 (continued)
One-tailed Sex with
Reference Age & Measure P advantage

Wispe, Kiecolt, & Long (1977)

Experiment 1 College Ratings of own mood in
(64) reaction to films of a
person being either
helped or not helped
Non-helping film
Aggression Significantf F
Sadness Significantf F
Elation ns NR
Helping film
Aggression ns NR
Sadness ns NR
Elation ns NR
Experiment 2 College Ratings of mood after
(64) viewing helping or non-
helping films (when
demand characteristics
were clear or unclear)
Aggression ns NR
Sadness ns NR
Elation ns NR
Zahn-Waxier, Friedman, & P, K, G I, G 2, Ratings of how sorry .015 F
Cummings (Note 13) G5,G6 subject felt for crying
(10 from each infant (tape recorded)
grade)
Zeskind & Lester (1978) College & Parents Ratings of reactions to ns* NR
(60) audiotape of crying
infant8

Note. P = preschool; K = kindergarten; G = grade; ns = not significant; but no p value was reported; NR = not
reported; M = male; F = female.
a
Personal communication from author.
b
Relevant ratings were "Sorry for," "distressed," "happy."
c
Relevant ratings were "happy," "indifferent," "distressed," "sympathetic."
d
Males reported more sympathy, although the main effect was not significant. Moreover, there was an interaction
of sex with type of cry and type of face (preterm/fultterm) for ratings of sympathy.
e
Sex differences were not reported, but means for both sexes were reported.
f
No probability value was reported.
B
The relevant rating was "distressed."
h
Only results significant at p < .01 were reported.

studies in which the children were observed influenced by the pattern of the children's
when they were not being explicitly ques- verbal responses to the stories. The fact that
tioned regarding their emotional responses, girls verbally reported more empathic re-
it seems likely that the sex differences in the sponses in the Marcus study is consistent
Sawin et al. and Marcus studies were related with this idea. (In Sawin et al. there was no
to the procedures being used. When children significant sex difference in verbal responses
are repeatedly questioned about their affec- to the PASTE, but the authors did not report
tive reactions to another's emotional state, which sex tended to respond more empath-
it is likely that many of them become aware. ically.)
that they are being evaluated, and try to re- Because specific data were reported in only
spond in a socially desirable way. Alterna- four studies, and two of these involved non-
tively, in studies in which children were ob- verbal responses while responding to the
served while responding to the PASTE, ob- PASTE, a meta-analysis was not computed.
servers' ratings may have been unconsciously Nonetheless, it is clear that researchers have
114 NANCY EISENBERG AND RANDY LENNON

Table 4
Sex Differences in Physiological Measures of Empathy
Sex with
Reference Age& N Measure One-tailed p advantage

Berger (1962)
Experiment 1 Adults GSR to seeing another shocked ns NR
(96)
Experiment 2 Adults ns NR
(32)
Craig and Lowery College Response to seeing another perform a
(1969) (56) task (and in some conditions being
shocked)
Heartrate NRa NR
GSR—vicarious instigation15 <.001C M
GSR—conditioned vicarious arousald .003 M
Frodi & Lamb (1978) 8-14 yrs. Reaction to cry of videotaped infant
(32) Heartrate ns NR
Skin conductance ns NR
Systolic & diastolic blood pressure ns NR
Frodi, Lamb, Lcavitt, Parents Reaction to crying of videotaped
& Donovan (1978) (56) infant
Skin conductance ns NR
Systolic & diastolic blood pressure .064 NR
Frodi, Lamb, Leavitt, Parents Reaction to crying of videotaped
Donovan, Neff & (64) newborns
Sherry (1978) Heartrate ns NR
Skin conductance ns NR
Systolic & diastolic blood pressure ns NR
Murray (1978; in press) College Reactions to audiotaped cried of
(64) newborns
Pulse amplitude ns NR
Pulse rate ns NR
Blood volume ?is NR
Stotland (1969, College Palmar sweating while observing NR M
reanalysis of (125) another taking a test"
Stotland & Dunn,
1963)
Stotland (1969, College Palmar sweating in response to NR M
reanalysis of (119) viewing another being tested5
Stotland & Walsh,
1963)

Note, ns = nonsignificant but no p level reported; NR = not reported; GSR = galvanic skin response; M = male;
F = female.
a
Since sex differences were reported for other measures and during the baseline period, it is likely that no sex
difference was found for this measure.
b
GSRs immediately after stimulus presentation.
c
On the basis of the discussion of the interaction between condition and sex, one can infer that males exhibited
more GSRs both when the confederate purportedly received shock and when he did not. Thus, this F (which
includes conditions in which the confederate was not shocked) is used in the table.
d
GSRs occurring 4 sec after stimulus presentation until stimulus onset.
e
The effect held not only for palmar sweating across conditions (when the person being tested did either poorly or
well), but also if one considers only the condition in which the testee did poorly.

found little evidence of a sex difference in served while responding to the story/slide
empathy when they have unobtrusively ob- measures of empathy (the PASTE), the pat-
served children's reactions to others' distress. tern of findings (a sex difference favoring fe-
In contrast, when children have been ob- males) parallels the results for research in
SEX DIFFERENCES IN EMPATHY 115

Tables
Sex Differences in Facial, Gestural, and Vocal (Tone of Voice) Measures of Empathy
Sex with
Reference Age & N Measure One-tailed p advantage
Hamilton (1973) P (24) Facial reactions to films ns NR
G 2 (24)
G 5 (24)
Leiman (Note 14) K(50) Facial reactions to videotape ns F
G 1 (35) of needy child
(N = 75 in this
analysis)
Lennon, Eisenberg, & 4-5 yrs. Facial/gestural reactions to
Carroll (Note 8) (35) videotape of distressed
children
Facial (Film 1) .317 M
Gestures (Film 1) .201 M
Latency to facial/gestural .230 F
response
Facial (Film 2) .086 M
Gestural (Film 2) .342 M
Latency to response .117 M
(Film 2)
Marcus (Note 9) P Children's reactions to
(32) PASTE stories
Facial measure ,293" F
Vocal measure .004" F
Sawin, Underwood, Weaver, 6-7 & 8-9 yrs. Reactions to PASTE stories
& Mostyn (Note 10) (90 in these
analyses) Facial measure .0003 F
Vocal measure .0027 F
Zahn-Waxler, Friedman, & P, K, G 1, Facial/gestural reactions to
Cummings (Note 13) G 2, G 5, tape of an infant crying
G6(10at Tension/anxiety .36a F
each age) measure a
Smiles measure . 18 Mb

Note. P = preschool; K = kindergarten; G = grade; ns = not significant but no p level reported; NR = not reported;
M = male; F = female. PASTE = Feshbach & Roe Affective Situations Test.
a
Personal communication from the author.
b
Females scored higher on smiling, which is consistent with the conclusion that females empatlu'zed less than males
with the distressed infant.

which empathy has been assessed with story/ Scale. As was mentioned previously, such
picture techniques. Thus, the pattern of data scales were designed to elicit information re-
is consistent with the conclusion that females garding the "trait" of empathy; that is, the
exhibit more nonverbal indices of empathy individual's enduring tendency to respond
(especially vocal affect) primarily when they empathically. Typical items in the Mehrabian
are also verbally reporting their empathic re- and Epstein scale include the following: "I
sponses, that is, when the fact that empathy tend to get emotionally involved with a
is being measured is relatively apparent. friend's problems," "Seeing people cry upsets
me," and "I tend to lose control when I am
Responses to Self-Report Scales bringing bad news to people." In studies with
children (e.g., Bryant, 1982; Kurtz & Eisen-
In studies with school children and adults, berg, in press), the wording of many items
empathy has frequently been assessed with has been simplified.
standardized pencil and paper question- The results of studies in which self-report
naires. The most commonly used of these is scales have been used are presented in Table
the Mehrabian and Epstein (1972) Empathy 6. As can be seen in this table, there is a huge
116 NANCY EISENBERG AND RANDY LENNON

sex difference in self-report of empathy as dren, was empathy assessed by means of peer
measured with questionnaires. In every study and/or teacher report (Barnett, Howard, Mel-
located, females obtained higher empathy ton, & Dino, 1982; Sawin et al., Note 10; see
scores. Moreover, with one exception (Hans- Table 7). In a manner consistent with cultural
son et al, in Stotland et al., 1978), the results stereotypes and the self-report studies, teach-
were significant. Especially for adults, the sex ers and peers viewed females as much more
differences were frequently so large that they empathic than males. Because only two stud-
were significant at higher than the p < ies were located, a meta-analysis was not
.000000001 level. Indeed, according to the computed. However, the findings in both
meta-analysis computed for these studies, the studies were highly significant.
finding of a sex difference favoring females
is Z = 18.35, p < .0001, with a mean effect Other Research Related to Sex Differences
size of .99 (see Footnote 1). According to the in Empathy
fail-safe statistic, 2,534 studies with a finding
of no sex differences would be needed to re- Some of the research related to sex differ-
duce the Z to below significance at the .05 ences in empathy has been conducted with
level. methods that do not neatly fit into any of the
Teacher- and peer-report measures. In various categories of research that we have
only two studies, both conducted with chil- used to organize our review (e.g., physiolog-

Table 6
Sex Differences in Self-Report of Empathy on Questionnaires

Sex with
Reference Age & N Measure One-tailed p advantage Effect size

Adams, Schvaneveldt, & G7(17), M&E <.001 F 1.75

Jenson (1979) G 8 (25),
G 9 (38)
Archer, Foushee, Davis, College M&E' NR b F
& Aderman (1979) (375)
Barnett, Feighy, & College M&E <.001 F .93
Esper (in press) (96)
C
Barnett, Howard, King, College M&E <.001 F .98
& Dino (1980) (295)
Barnett, King, Howard, Parents of M&E <.001 F 1.33
& Dino (1980) preschoolers
(108)
Barnett, Howard, King, G 1 0 & G 12 M&E •e.OOl F 1.35
& Dino (1981) (103)
Bryant (1982) G 1 (56), Modified <.001 F .80
G4(115), M&E
G 7 (87)
G 7 only M&E <.001d F 1.07
Eisenberg-Berg & G 9 (25), M&E •c.OOl F 1.61
Mussen (1978) G 1 1 (22),
G 12 (25)
Foushee, Davis, & College M&Ea <.001 F .97
Archer (1979) (591)
Hansson, Chernovetz, & College Stotland et al., .139 F .13
Jones (unpublished, (179) Fantasy-
reported in Stotland Empathy
etal., 1978) Scale
Kalliopuska(1980) 9-12 yrs. M&E <.001 F .77
(333) (3 pt. scales)
Finnish
sample
Parents of 9- M&E <.001£ .65
12 yr. olds
(341)
SEX DIFFERENCES IN EMPATHY 117

Table 6 (continued)
Sex with
Reference Age&W Measure One-tailed p advantage Effect size

Kurtz & Eisenberg 03 M&E .002 F .50

(in press) (129) (4 pt. scales)
Mehrabian (1977) College M&E <.001 F 1.71
(325)
f
Mehrabian & Epstein College M&E .025 F .43
(1972) (88)
Mehrabian & Ksionsky College M&E <.001 F 1.38
(1974)8 (196)
Stotland, Mathews, College Stotland et al.'s
Sherman, Hansson, (516) self-report
& Richardson scales:
(1978)
Denial- <.001 F"
avoidance
Involvement- <.001 F
concern
Hostility- .006 Fh .48
empathy
Friend- •c.OOl F
empathy
Fantasy- .012 F
empathy

Note. G = Grade; M & E = Mehrabian and Epstein Empathy Scale; ns = nonsignificant but no p level reported;
F = female; M = male; NR = not reported.
" A few items of the scale were dropped.
b
Although statistical tests were not reported, the sex difference in median empathy scores was large.
c
For the subsample of subjects chosen to participate in the study, /(70) = 2.58, p < .05.
d
Personal communication from the author.
e
Computed from means and standard deviations.
f
No F value was given although the sex difference was reported to be significant; thus, a conservative estimation
of p < .05 was used in the meta-analysis.
B
This is apparently a reanalysis of some data presented in Mehrabian and Epstein (1972), in which six subjects
were omitted. Thus, those data in Mehrabian and Epstein (1972) are not included in this table.
h
Females scored lower on this scale, indicating that they were more empathic.

ical measures, reflexive crying). This research Zahn-Waxler trained mothers of children
will now be discussed. aged 10 to 20 months to observe their own
Two excellent studies that contain relevant children and to report in detail the ways in
data are those conducted by Baumrind (1971, which the children responded to emotion-
1980, Note 15) and by Radke-Yarrow and evoking incidents (especially incidents in-
Zahn-Waxier (in press). Radke-Yarrow and volving another's distress). The children were

Table?
Sex Differences in Other-Report Measures of Empathy
Sex with
Reference Age & A^ Measure One-tailed p advantage

Barnett, Howard, Melton, G6 Peer ratings <.001a F

& Dino (1982) (112)
Teachers' ratings <.001" F
Sawin, Underwood, Weaver, 6-7 & 8-9 yrs. Teachers' ratings .006 F
& Mostyn(Note 10) (128)

Note. G = grade; F = female.

a
Personal communication from author.
118 NANCY EISENBERG AND RANDY LENNON

also observed by the researchers as they re- appreciates another child's point of view,
sponded to others' simulated affect (simu- does not assault another child's ego, respects
lated by mothers and/or experimenters). Ob- other children's work, trustworthy; Baum-
servations were made over a 9-month period, rind, Note 16). Thus, although girls scored
and the children were studied for three ad- higher on the friendly-hostile cluster (exhib-
ditional months at 6-7 years of age. ited more positive behaviors) at both ages 4
On the basis of data described above, and 9, it is unclear if the girls were really
Radke-Yarrow and Zahn-Waxler classified more empathic. Nonetheless, Baumrind's
the ways in which the children responded to data are certainly consistent with the conclu-
others' distress. Some children exhibited clear sion that girls are more oriented to others'
empathic responses while viewing another's needs than are boys.
sadness or distress. Others exhibited more Another type of data that is relevant to the
cognitive, combative, problem-solving, anx- discussion of sex differences in empathy con-
ious and "guilty," or intolerant responses. cerns moral reasoning about prosocial moral
These styles of reacting were relatively stable conflicts. Eisenberg and her students have
across the 5- to 6-year period between the conducted a number of studies in which chil-
initial data collection and the follow-up. dren and adolescents were interviewed about
Moreover, and most relevant to the present helping and sharing situations. More specif-
discussion, there were no sex differences in ically, children were interviewed about either
style of reacting to others' distresses. How- their own naturally occurring positive be-
ever, the number of subjects in this research haviors (Eisenberg-Berg & Neal, 1979), or
was relatively small; sex differences may have hypothetical moral dilemmas in which the
been evident if the sample had been larger. story protagonist could assist another, but at
Baumrind (1971, 1980, Note 15), like a cost to the self (Eisenberg, 1977; Eisenberg-
Radke-Yarrow and Zahn-Waxler, has been Berg, 1979; Eisenberg-Berg & Roth, 1980;
conducting longitudinal research related to Eisenberg-Berg & Neal, 1981; Eisenberg,
the development of positive (and negative) Lennon, & Roth, in press).
behaviors in children. The preschool children In this research, the children's reasoning
(aged 4) in this study were observed in nurs- was coded into a number of different cate-
ery schools for approximately 30 hours, and gories, some of which reflected empathic con-
their parents were interviewed at home. The cerns. For example, in research with pre-
same children were studied at age 9. schoolers and children in early elementary
In general, Baumrind (1980, Note 15) school, children have frequently oriented di-
noted that girls were more prosocial and em- rectly toward another's needs in discussing
pathic than boys. Unfortunately, however, it a proposed course of action. In most of the
is impossible to tell whether the girls were research, no sex differences in empathic rea-
really more empathic, that is, were explicitly soning have been found (Eisenberg-Berg,
more affectively responsive and/or sympa- 1979; Eisenberg-Berg & Hand, 1979; Eisen-
thetic to another's affective state. This is be- berg-Berg & Neal, 1979, 1981; Eisenberg,
cause the data related specifically to empathy Lennon, & Roth, in press). The only excep-
were included in factor groupings (based on tion was in Eisenberg (1977). In this study
the results of a factor analysis) with data re- involving 2nd, 4th, 6th, 9th, 11th, and 12th
lated to other characteristics of the children. graders, girls expressed significantly more
For example, in the cluster of items called concern with others' psychological needs, as
"hostile-friendly" were items related to em- well as more sympathetic reasoning. How-
pathy ("sympathetic"), affective role taking ever, there were no sex differences in the use
("understands other children's position in in- of several other categories of reasoning re-
teraction"), and prosocial/antisocial behavior lating to empathy or role taking, and girls did
(e.g., "bullies other children," "selfish"; 1971, not exhibit more use of empathic/sympa-
p. 7). At age 9, the hostile-friendly cluster thetic reasoning when the sexes were com-
also contained items related to both empathy pared on their overall, modal "stage" of rea-
and other behaviors (sympathetic, altruistic, soning (where stages were constructed by
SEX DIFFERENCES IN EMPATHY 119

.combining various types of reasoning that found little evidence of a sex difference. Con-
clustered together, such as various categories sequently, he concluded that sex differences
of empathic and sympathetic reasoning). in empathy were not due to differences in
In summary, the research of Radke-Yar- males' and females' affective role taking
row and Zahn-Waxier, Baumrind, and Eisen- skills.
berg provides relatively little supporting evi- In his review of the research, Hoffman fo-
dence of clear sex differences in the tendency cused primarily on the literature that tradi-
to empathize. Although Baumrind found tionally has been associated with affective
that girls are more prosocial than boys, it is role taking; that is, research measuring re-
unclear whether the girls in her studies were actions to pictures and/or stories. For ex-
also more empathic than the boys. Moreover, ample, with one popular measure of affective
Radke-Yarrow and Zahn-Waxier and Eisen- role taking (Borke, 1971, 1973), subjects are
berg found no evidence of consistent sex dif- told a series of brief stories about events that
ferences either in mode of responding to an- are likely to make people either happy, sad,
other's distress or in reasoning about helping afraid, or angry. After each story, they are
dilemmas. questioned about the story protagonist's af-
fect. Subjects indicate the character's affect
Summary of the Empathy Research by selecting one of four stylized pictures of
faces depicting different emotions. Other
In summary, the results of research con- measures of affective role taking (such as
cerning sex differences in empathy are highly Feshbach & Roe's, 1968, comprehension
related to the method of assessing empathy. measure) involve the use of pictures to illus-
Sex differences favoring females have been trate the hypothetical situation and fre-
strongest when self-report or other-report quently include pictures of the story protag-
measures were used to assess empathy. Fe- onist's facial expression. However, despite
male infants also appear to exhibit somewhat differences in methods, in all the "tradi-
more reflexive crying than males. Few sex tional" affective role taking tasks, some in-
differences have been found when empathy formation regarding the story protagonist's
was assessed with nonobtrusive measures, for situation is always presented; that is, in these
example, physiological or facial/gestural tasks, facial expressions or other nonverbal
measures. cues are never the only information provided
Sex Differences in Affective Role Taking for the subjects.
Hall (1978) noted that Hoffman, in his re-
Although the ability to recognize another's view, omitted much of a relatively large body
emotional state (affective role taking) may of research on affective role taking—the re-
not be a prerequisite of all types of empathiz- search on sex differences in decoding of non-
ing (Hoffman, 1982, Note 1), affective role verbal cues; that is, body cues, facial expres-
taking is believed to be a component of many sions, and tone of voice. She reviewed 75 em-
forms of empathy, especially more mature pirical studies of the ability to interpret facial
forms of empathizing (Feshbach, 1978; Hoff- expressions, body cues, and/or auditory cues
man, 1976, 1982). Because of the intimate related to another's affective state. On the
theoretical association between affective role basis of her review and a meta-analysis, Hall
taking and empathy, both theorists and re- concluded that females are better able than
searchers have frequently discussed and/or males to decode visual, as well as auditory,
examined the role of affective role taking in cues.
empathy. If there is, indeed, a sex difference in de-
Hoffman (1977) reviewed the research on coding ability, it would be reasonable to hy-
affective role taking to determine whether sex pothesize that any sex differences in empathy
differences in role taking might explain the might be due, in part, to the tendency for
sex differences he noted in empathic respond- females to be superior decoders of others'
ing. He summarized 12 reports related to affective states. Certainly, if girls and women
affective role taking, all with children, and are better able than males to ascertain an-
120 NANCY EISENBERG AND RANDY LENNON

other's affective state from nonverbal cues, visual decoding skills of children are consid-
one would expect females to react to others' erably less convincing.
emotional states more frequently than males. In Table 8, we have summarized the stud-
For this reason, discussion of sex differences ies concerning children's visual decoding
in affective role taking was viewed as relevant skills, including the research reviewed by Hall
to this review. as well as several more recent studies. In the
The literature concerning affective role 15 references reviewed (involving 19 sam-
taking skills is organized by method into two ples), females scored significantly or margin-
groupings: research on decoding skills, and ally significantly higher in only two studies,
research involving picture/story techniques. on two of four measures in one study (Staf-
Because it is difficult to determine the psy- fieri & Bassett, 1970), and on one of two
chological significance of variations in in- measures for each of two different samples
fants' discriminations of different facial in another study (Blanck et al., 1981). In two
expressions (which are usually measured by other studies (Dickey & Knower, 1941; Kel-
differential fixation times) the research on logg & Eagleson, 1931), females scored some-
infants' reactions to facial expressions will what higher, but statistical tests of sex differ-
not be reviewed in this paper. It is interesting ences were not computed. In contrast, males
to note, however, that researchers have found scored significantly higher on one of four
few sex differences in infants' reactions to measures in one study (Staffieri & Bassett,
different facial expressions (e.g., Barrerra & 1970). Moreover, boys scored higher in the
Maurer, 1981; LaBarbera, Izard, Vietze, & preschool years in one study in which girls
Parisi, 1976; Oster, 1981; Oster & Ewy, re- excelled in decoding in the school years (Kel-
ported in Oster, 1981; Wilcox & Clayton, logg & Eagleson, 1931, no statistics reported).
1968; Young-Browne, Rosenfield, & Horow- In the remaining 11 studies, there was little
itz, 1977). evidence of a sex difference.
According to our review of the empirical
Nonverbal Decoding data, there is little evidence of a consistent
sex difference in children's decoding of visual
As was mentioned previously, Hall (1978) nonverbal cues. A meta-analysis was not
published an extensive review on sex differ- computed to confirm this conclusion because
ences in decoding. She found that females the necessary data were not included in most
were better visual and auditory decoders than of the research reports.
males, and that this sex difference was not The studies reviewed by Hall (1978) and
a function of age. those presented in Table 8 concern decoding
It is not our goal to duplicate Hall's work; of overt (nondeceptive) visual cues. In the
however, we disagree, in part, with her con- literature related to the decoding of deceptive
clusions. With regard to the decoding skills nonverbal cues, sex differences in decoding
of adults, our reading of the data is similar are even less evident than in the research con-
to Hall's. In our opinion, the empirical data cerning the decoding of overt, nondeceptive
are consistent with Hall's conclusion that cues. According to a recent review on the
women are more skilled than men at visually communication of deception (Zuckerman,
decoding others' nonverbal cues (although, DePaulo, & Rosenthal, 1981), females are
because we assigned a finding of "no differ- slightly better detectors of deception than are
ence" an effect size of 0 and Hall did not, we males (mean d = .17), but the difference is
do not agree that the finding is as strong as small and nonreliable. This absence of con-
asserted by Hall). Moreover, according to the sistent sex differences is evident in the liter-
limited data reviewed by Hall and more re- ature involving children (DePaulo, Jordan,
cent research (Blanck, Rosenthal, Snodgrass, Irvine, & Laser, 1982; White & Feldman,
DePaulo, & Zuckerman, 1981; Rosenthal Note 19), as well as adults (Zuckerman et al.,
& DePaulo, 1979), it appears that females are 1981). More specifically, as nonverbal cues
indeed better decoders of auditory cues than become less intended, females show a de-
are males, both in childhood and adulthood. creasing advantage over males in the accu-
However, Hall's conclusions regarding the racy of decoding (Rosenthal & DePaulo,
SEX DIFFERENCES IN EMPATHY 121

Table 8
Sex Differences in Children's Decoding of Visual Nonverbal Cues
Sex with
Reference Age &N Measure One-tailed p advantage
Blanck, Rosenthal, Snodgrass,
DePaulo, & Zuckerman
(1981)"
Group 1 9-15 yrs. Films: faces .319 M
(250)
Films: body .001 F
movement
Group 2 High school Films: faces .033 F
(109)
Films: body .109 F
movement
Buck (1976) High school Photos: faces ns NR
(10)
Ekman&Friesen(1971) New Guinea Videotapes: faces ns F
children
(130)
Dickey & Knower Jr. high Photos: faces NR F
(1941) (233 Americans;
70 Mexicans)
Felleman, Harden, Carlson, P Photos: faces ns NR
Rosenberg, & Masters (20)
(Note 17)
Gates (1923) 3-14 yrs. Photos: faces ns NRb
(458)
Gitter, Mostofsky, & Quincey 4-6 yrs. Slides: faces ns NR
(1971) (80)
Hamilton (1973) P (24), Matching photographs ns NR
G 2 (24), of facial expressions
G 5 (24)
Hebda, Peterson, & Miller G3(3I) Pictures of faces ns NR
(1972)° (rated on anger-
friendliness
dimension)
Krauss & Morency P(19) Videotapes: faces ns NR
(Note 18)
Kellogg & Eagleson 3-14 yrs. Pictures: faces NR F"
(1931) (332)
Odom & Lemond (1972) K(16), Matching photographs ns NR
G 5 (16) of similar facial
expressions
Rosenthal, Hall, DiMatteo, G4&G 5 Films: face and body IIS F
Rogers, & Archer ( 1 979; (275)
summarized in Hall,
1978)
Staffieri & Bassett (1970) 10-11 yrs. Pictures: faces
Pictures of:
Group 1 (60) adult female .014 M
Group 2 (60) adult male .068e F
Group 3 (49) child female .011 F
Group 4 (59) child male NR M
Zuckerman & Przewuzman 2'/2-5 yrs. Slides and photo- ns NR
(1979) (77) graphs: faces

Note. P = preschool; K = kindergarten; G = grade; ns = not significant, but no p value was reported; NR = not
reported; F = female; M = male.
"The high school subjects are the same as in Rosenthal and DePaulo (1979).
b
No tests of significance were reported, but sex differences were slight and inconsistent (with the exception of boys'
consistent superiority in the labeling of fear).
c
Although Hall (1978) did not view this as a study of decoding, we believe this study to be relevant.
d
Only Ms and Ns were given. We knew of no way to compute a p value, although Hall (1978) apparently did
compute effect size.
° The t value and p value were inconsistent, so the probability level was recomputed from the t value.
122 NANCY EISENBERG AND RANDY LENNON

1979). Moreover, according to recent re- could not be computed. Nonetheless, it is

search with subjects 9 years and older, with clear that there is no consistent sex difference
increasing age females lose more and more in affective role taking as assessed with tra-
of their advantage for decoding covert non- ditional picture/story techniques.
verbal cues, but gain more and more of an
advantage for decoding overt cues (Blanck et Summary of the Research on Affective Role
at., 1981; Zuckerman, Blanck, DePaulo, & Taking and Decoding
Rosenthal, 1980). To summarize, although women appear to
be more skilled than men at decoding visual,
Traditional Tests of Affective Role Taking overt nonverbal cues (Hall, 1978), there seem
to be relatively few consistent sex differences
As can be seen in Table 9, sex differences in children's affective role taking (including
have been found in very few studies involving decoding) abilities. Girls do seem to be better
picture/story measures of affective role tak- than boys at decoding auditory cues; how-
ing. In only 3 of 23 research reports did fe- ever, there is relatively little evidence that girls
males score significantly higher on affective and boys differ in ability to decode visual
role taking (Borke, 1973; Hudson, 1978; nonverbal cues or in performance on tradi-
Staub & Feinberg, Note 5), while boys scored tional affective role taking tasks. Moreover,
significantly higher in one study (Kurdek & with increasing age, males tend to become
Rodgon, 1975). more skilled than females at decoding decep-
Because most researchers who conducted tive or covert cues.
the studies in Table 9 merely reported a find- Based on the pattern of findings in the lit-
ing of "no sex difference," a meta-analysis erature on decoding and affective role taking,

Table 9
Sex Differences in Performance of Traditional Affective Role Taking Tests
Sex with
Reference Age& N Measure One-tailed p advantage
Borke (1971) 3-8 yrs. NVR ns NR
(200)
Borke (1973) 3-6 yrs. NVR <.001 F
(288 Chinese;
288 American)
Brehm, Powell, & Coke G 1 (67) VR ns NR
(in press)
Deutsch (1975) 3-4 yrs. VR & NVR ns NR
(32)
Eisenberg-Berg & 4-5 yrs. VR .380 M
Lennon (1980) (51)
NVR .186 M
Feshbach & Feshbach 4-5 yrs. VR ns F
(1969) (48)
6-7 yrs. VR ns M
(40)
Feshbach & Roe G 1 (46) VR ns Ma
(1968)
Hudson (1978) G2 (110) VR
Angry ns NR
Sad ns NR
Happy .024 F
Hughes, Tingle, & K&G2 VR ns NR
Sawin(1981) (48)
lannotti (Note 3) 4-5 yrs. NVR
(52) Empathy based on situational M
cues
Empathy based on facial F
cues
SEX DIFFERENCES IN EMPATHY 123

Table 9 (continued)
Sex with
Reference Age &N Measure One-tailed p advantage

Kurdek & Rodgon K-G6 VR

(1975) (167) Projection of inappropriate .001 Mb
responses
Labeling of affect when <.001 M
character's affect was
different from
situational cues
Labeling affect when it was us M
consistent with
situational cues
Lennon, Eisenberg, & P VR .324 M
Carroll (Note 7) (80)
Levine & Hoffman 4 yrs VR ns F
(1975; reported in (77)
Hoffman, 1977)
Marsh, Serafica, & G8 VR us NR
Barenboim (1980, (64)
1981)
Mood, Johnson, & P(40) VR ns NR
Shantz (1978)
Odom & Lemond K(16), NVR ns NR
(1972) G 5 (16)
Radin & Sagi 3-6 yrs, NVR .383 NR
(1982) (59 Americans;
60 Israelis) .172 NR
Rothenberg G 3 (57), VR .367 F
(1970) 05(51)
Rubin (1978) P (26), NV ns NR
G 1 (41)
Saarni (1979) 6 yrs. (20) Complexity of reasoning ns Fc
8 yrs. (20) about affect ns F
10 yrs. (20) ns M
Staub & Feinberg G3-4 V <.001d F
(Note 5) (51)c
Urberg & Docherty 3 yrs. (14), VR ns NR
(1976) 4 yrs. (14),
5 yrs. (14)
Walton (1936) 6-16 yrs. (693) VR ns NR
College (131) ns NR

Note. P = preschool; K = kindergarten; G = grade; NVR = nonverbal (pointing) labeling of story protagonist's
affect; VR = verbal labeling of story protagonist's affect; NS = not significant; but no p level reported; NR = not
reported; F = female; M = male.
a
Computed from means. b Males project less, indicating higher role taking. c At age 10, but not 6 or 8, boys tended
to score higher. d Personal communication from author.

a reasonable conclusion is that any sex dif- dures in which the subject hears only a bland,
ferences in adults' empathy could be due, in unemotional description of the story protag-
part, to sex differences in role taking and de- onist's situation, the sex difference in audi-
coding skills. However, except for instances tory decoding could not account for most
in which affective cues were communicated sex differences in research on children's
auditorially, it would be difficult to attribute empathy.
any sex differences in children's empathy to
sex differences in either affective role talcing Discussion
or decoding abilities. Moreover, because most
of the research on children's empathy has It is clear from the review of the empirical
been conducted with picture/story proce- literature that the data regarding sex differ-
124 NANCY EISENBERG AND RANDY LENNON

ences in empathy are inconsistent, and that when it is obvious what behavior or trait is
this inconsistency is a function of the method being assessed. Thus, when individuals have
used to measure empathy. Sex differences in been asked to rate themselves on behaviors
empathy favoring females are most evident or reactions clearly related to the concept of
when individuals have been asked to rate empathy, females have scored much higher
themselves on behaviors or affective respon- on empathy than have males. When the de-
ses related to the concept of empathy and/or mand characteristics inherent in a method-
sympathy. Somewhat weaker sex differences ology have been a bit more subtle, that is,
favoring females have been found when sub- when the purpose of the assessment situation
jects have been asked to rate their emotional was not entirely clear, females have still
responses in contrived situations or in re- tended to score higher on empathy, but the
sponse to hypothetical picture/story mea- sex difference is much smaller. Finally, when
sures. In contrast, few consistent sex differ- the measure of empathy has been even more
ences in empathy have been noted in research unobtrusive, for example, when physiological
in which empathy was assessed with physi- measures or facial/gestural measures have
ological measures and/or facial/gestural mea- been the indices of empathy, females have
sures (except when the child was being tested not exhibited more empathy than have males.
with a picture/story measure at the same time The only deviation from the pattern of as-
he or she was being observed). sociation between degree of demand char-
How can this inconsistent pattern of re- acteristics in the assessment situation and the
sults be explained? One explanation might size of the sex difference in empathy is for
be related to age of the subjects used in the data on reflexive crying; however, as was dis-
different types of research. Sex differences cussed previously, the significance of reflexive
were most evident in research involving self- crying for an understanding of later sex dif-
report questionnaires. This research gener- ferences in empathy is not clear.
ally has been conducted with older children Because the size of the sex difference in
and adults. Moreover, consistent sex differ- empathy varies greatly across methods, ques-
ences in decoding skills, which could engen- tions about the validity of the different mea-
der sex differences in empathic responding, surement instruments naturally arise. Con-
seem to occur mostly in adulthood. Thus, sidering just the issue of demand character-
one could hypothesize that sex differences in istics, it is reasonable to hypothesize that
empathy increase with age. However, such an methods that are more unobtrusive, for ex-
explanation is not viable because there is little ample, facial/gestural and physiological mea-
evidence of a sex difference in empathy when sures, are more valid indices of empathy than
adults' empathic responses have been mea- are measures of empathy involving more de-
sured physiologically. Moreover, in the stud- mand characteristics. In fact, according to the
ies in which adults have reported their affec- limited data, there is little relation between
tive responses to simulated empathy-provok- empathy as measured with self-report mea-
ing situations, there is much less evidence of sures and with less obtrusive measures (Craig
a sex difference favoring females than when & Lowery, 1969; Zahn-Waxier et al., Note
empathy has been assessed with question- 13). Moreover, empathy as assessed with the
naires. Finally, when empathy has been op- PASTE is unrelated (Lennon et al., Note 8) or
erationalized as either reflexive crying or as inconsistently related (Sawin et al., Note 10)
a matching response on picture/story mea- to facial/gestural/vocal measures. Clearly,
sures, relatively small but significant sex dif- there is reason to doubt that all measures of
ferences in children's empathy have been empathy are equally valid.
found. Unfortunately, because empathic respon-
Given the discrepancies in the findings ses are primarily an internal reaction, it is
across the different methods used to assess difficult to determine the validity of any given
empathy, it would be logical to hypothesize measure of empathy. Nonetheless, because
that the inconsistent pattern of results is due of the theoretical association between em-
to differences in the various methods. The pathy and altruism (Hoffman, 1981, 1982;
sex difference in empathy is most evident Mussen & Eisenberg-Berg, 1977; Staub,
SEX DIFFERENCES IN EMPATHY 125

1978), researchers sometimes have attempted but only when there were clear demand char-
to validate a measure of empathy by exam- acteristics in the experimental situation.
ining the relation between empathy scores There certainly is reason to believe that
obtained with a specific measure and indices males and females might differ in how em-
of prosocial and/or antisocial behaviors (e.g., pathic they would like to appear to others
Feshbach & Feshbach, 1969; Mehrabian & (and, perhaps, to themselves). Emotionality
Epstein, 1972; Stotland, Mathews, Sherman, and nurturance are both part of the stereo-
Hansson, & Richardson, 1978). If an asso- typic feminine role (e.g., Block, 1973). Thus,
ciation between a measure of empathy and it is highly likely that females would be more
prosocial behavior is considered an accept- willing than males to present themselves as
able criterion for external validity, several of being empathic and/or sympathetic, even if
the measures of empathy appear to be at least there were no real sex difference in respon-
somewhat valid. Scores on self-report mea- siveness. Indeed, self-ratings of adults' femi-
sures such as Mehrabian and Epstein's have ninity have been positively related to men's
frequently been associated with prosocial be- and women's self-report of empathy, whereas
havior (Archer, Diaz-Loving, Gollwitzer, reports of masculinity have been negatively
Davis, & Foushee, 1981; Barnett, Howard, associated with empathy scores (Foushee,
King, & Dino, 1981; Eisenberg-Berg & Mus- Davis, & Archer, 1979). Self-presentational
sen, 1978; Liebhart, 1972; Mehrabian & Ep- and self-esteem issues would be expected to
stein, 1972), as have self-report measures in be especially salient when people are asked
experimental settings (Batson et al., Note 12; to rate themselves on items that clearly reflect
Marks, Penner, & Stone, 1982), other-report the feminine traits of expressivity, sentimen-
measures (Barnett et al., 1982), physiological tality, and sympathy (as on Mehrabian and
measures (Krebs, 1975), and facial/gestural Epstein's scale). Consequently, it is not sur-
measures (Leiman, Note 14; Lennon et al., prising that a sex difference in empathy is
Note 8; Peraino & Sawin, Note 20; but not found primarily when subjects are asked to
Zahn-Waxler et al, Note 13). In contrast, rate themselves on traits that are clearly mas-
empathy as measured by the PASTE and de- culine or feminine.
rivative instruments has only occasionally Research on behaviors other than empathy
been associated with prosocial behavior, is consistent with the hypothesis that sex dif-
(Feshbach, 1982; Marcus, Telleen, & Roke, ferences with regard to sex-typed behaviors
1979) and has more often been negatively are a function, in part, of the clarity of sex-
related, inconsistently related, or unrelated role expectations. Berman (1980), in her re-
to prosocial behavior (Eisenberg-Berg & Len- view of the literature related to sex differ-
non, 1980; Fay, 1971; lannotti, 1978, Note ences in responsiveness to the young, found
21; Lennon et al., Note 8; Sawin et al., Note that males and females differed greatly in self-
11; c.f. Underwood & Moore, 1982). In brief, report of their reactions to infants. However,
with the possible exception of measures sim- she found no sex difference in physiological
ilar to the PASTE, the external validity of most responding to infants and children; and fe-
types of empathy measures has received sup- males exhibited more behavioral responses
port from the empirical literature on proso- toward the young than did males, primarily
cial behavior. in situations in which they were explicitly or
The idea that self-report of empathy may implicitly assigned to the care-taking role. In
be influenced by demand characteristics has brief, Berman's findings can be interpreted
received some empirical support. Wispe et as indicating that there are sex differences in
al., (1977) varied the explicitness of the de- responsiveness to the young (a feminine be-
mand characteristics in a situation in which havior), but only in situations in which (a)
adult subjects were asked to rate their own it is clear that subjects are being evaluated on
mood after viewing films that would be ex- that dimension, or (b) subjects are in circum-
pected to elicit either a positive or negative stances in which role expectations or obli-
empathic response. They found a strong re- gations are salient.
lation between the outcome for the story pro- Another possible explanation for the dis-
tagonist in the film and self-report of affect, crepancies in the research concerning em-
126 NANCY EISENBERG AND RANDY LENNON

pathy is that an empathic reaction may not Given the inconsistent pattern of data on
always be expressed simultaneously via all sex differences in empathy, the conclusions
possible channels (e.g., self-report, facial and that can be drawn from the current empirical
physiological reactions), and different people literature must be circumscribed and tenta-
may tend to express empathy in different tive. Certainly, it is true that self-report of
ways. Buck and his colleagues have found empathic or sympathetic reactions is stronger
that children (Buck, 1977) and adults (Buck, for females than for males. However, it also
Miller, & Caul, 1974; Buck, Savin, Miller, is clear that females' reputations for and self-
& Caul, 1972) who clearly communicate report of empathic responding are much
their emotional responses via facial expres- stronger than are their tendencies to respond
sions tend to exhibit relatively few physio- (physiologically or nonverbally) in an em-
logical reactions in the same situation. Buck pathic manner. As sex roles in Western cul-
has labeled people who exhibit emotional re- ture become less distinct (as seems to be hap-
actions externally rather than internally (fa- pening), the large sex difference in self-report
cially rather than physiologically) as exter- and other-report of empathy and sympathy
nalizers, and people who exhibit little exter- may diminish in size. Of course, if the sex
nal affect but respond physiologically as difference in empathy is due, in part, to in-
internalizers. Although not all researchers nate differences in the capacity for empathy,
have found a negative relation between an one could not expect the sex difference in
individual's physiological responding and the empathy to disappear even if the culture did
accuracy of his or her facial communication change dramatically. However, the limited
(Zuckerman, Klorman, Larrance, & Spiegel, empirical research, at this time, provides little
1981), it appears that for a given individual, basis for the conclusion that the sex differ-
one type of emotional response may fre- ence in empathy found for self-report mea-
quently predominate over and perhaps in- sures is due to an innate mechanism or pre-
hibit alternative modes of response. This phe- disposition. Indeed, at the present, all that
nomenon may explain why various measures can be concluded with confidence is that
of an individual's empathic reaction to a spe- many important issues concerning sex dif-
cific situation are frequently unrelated. ferences in emotional empathy are, as yet,
Buck found a sex difference in the ten- unresolved.
dency to be either an externalizer or an in-
ternalizer; women tend to be externalizers Reference Notes
and men tend to be internalizers (Buck, 1981;
Buck et al., 1972, 1974). If verbalization of 1. Hoffman, M. L. (Discussant). Symposium on "Em-
affect is considered a method of externalizing pathy in children: Conceptual and methodological
affect, this sex difference in mode of respond- issues in current research." Presented at the Biennial
Meeting of the Society for Research in Child De-
ing to emotion-provoking stimuli could ex- velopment, San Francisco, March 1979.
plain, in part, the fact that there is a strong 2. Roe, K. Personal communication, March 1982.
sex difference for verbal measures (males 3. lannotti, R. J. Prosocial behavior, perspective taking,
would be expected to score higher on phys- and empathy in preschool children: An evaluation of
naturalistic and structural settings. Paper presented
iological indices of empathy). However, the at the Biennial Meeting of the Society for Research
internal/external differentiation cannot ac- in Child Development, Boston, April 1981.
count for the discrepancies across methods 4. Miller, S. M. Dependency, empathy, and altruism.
in data relating to children's empathy for two Paper presented at the Biennial Meeting of the So-
ciety for Research in Child Development, New Or-
reasons: (a) the internalizing/externalizing leans, March 1977.
sex difference does not appear to exist for 5. Staub, E., & Feinberg, H. K. Regularities in peer
children, at least for preschoolers (Buck, interaction, empathy, and sensitivity to others. Paper
1977), and (b) there are few sex differences presented at the Annual Meeting of the American
in empathy as measured with either unob- Psychological Association, Montreal, September
1980.
trusive observations of facial/gestural (exter- 6. Eisenberg-Berg, N., & Lennon, R. Unpublished
nal) responses or physiological (internal) data, Arizona State University, 1980.
measures. 7. Lennon, R., Eisenberg, N., & Carroll, J. The effects
SEX DIFFERENCES IN EMPATHY 127

of sex of experimenter and sex of subject on gender Archer, R. L., Diaz-Loving, R., Gollwitzer, P. M., Davis,
differences in preschool children's empathy. Unpub- M. H., & Foushee, H. C. The role of dispositional
lished manuscript, Arizona State University, 1982. empathy and social evaluation in the empathic me-
8. Lennon, R., Eisenberg, N., & Carroll, J. The relation diation of helping. Journal of Personality and Social
between empathy and prosocial behavior in the pre- Psychology, 1981, 40, .786-796.
school years. Unpublished manuscript, Arizona Archer, R. L., Foushee, H. C., Davis, M. H., & Aderman,
State University, 1982. D. Emotional empathy in a courtroom simulation: A
• 9. Marcus, R. Unpublished data, University of Mary- person-situation interaction. Journal of Applied Social
land, 1982. Psychology, 1979, 3, 275-291.
10. Sawin, D. B., Underwood, B., Weaver, J., & Mostyn, Aronfreed, J. The socialization of altruistic and sym-
M. Empathy and altruism. Unpublished manu- pathetic behavior: Some theoretical and experimental
script, University of Texas, Austin, 1981. analyses. In J. Macaulay & L. Berkowitz (Eds.), Al-
11. Sawin, D. Assessing empathy in children: A search truism and helping behavior. New York: Academic
for an elusive concept. Paper presented at the Bien- Press, 1970.
nial Meeting of the Society for Research in Child Bakan, D. The quality of human existence. Boston: Bea-
Development, San Francisco, March 1979. con Press, 1966.
12. Batson, C. D., O'Quin, K., Fultz, J., Vanderplus, M., Barnett, M. A., Feighny, K. M., & Esper, J. A. Effects
& [sen, A. M. Influence of self-reported distress and of anticipated victim responsiveness and empathy
empathy on egoistic versus altruistic motivation to upon volunteering. Journal of Social Psychology, in
help. Unpublished manuscript, University of Kan- press.
sas, 1982. Barnett, M. A., Howard, J. A., King, L. M., & Dino,
13. Zahn-Waxier, C, Friedman, S. L., & Cummings, G. A. Antecedents of empathy: Retrospective ac-
E. M. The development of helping behavior: Chil- counts of early socialization. Personality and Social
dren's responses to infants' cries. Unpublished Psychology Bulletin, 1980, 6, 361-365. (a)
manuscript, Laboratory of Developmental Psychol- Barnett, M. A., King, L. M., Howard, J. A., & Dino,
ogy, National Institutes of Mental Health, 1982. G. A. Empathy in young children: Relation to parents'
14. Leiman, B. Affective empathy and subsequent altru- empathy, affection, and emphasis on the feelings of
ism in kindergarteners and first graders. Paper pre- others. Developmental Psychology, 1980,16, 243-244.
sented at the annual meeting of the American Psy- (b)
chological Association, Toronto, August 1978. Barnett, M. A., Howard, J. A., King, L. M., & Dino,
15. Baumrind, D. Sex-related socialization effects. Pa- G, A, Helping behavior and the transfer of empathy.
per presented at the Biennial Meeting of the Society Journal of Social Psychology, 1981, 115, 125-132.
for Research in Child Development, San Francisco, Barnett, M. A., Howard, J. A., Melton, E. M., & Dino,
March 1979. G. A. Effect of inducing sadness about self or other
16. Baumrind, D. Personal communication, March 30, on helping behavior in high and low empathic chil-
1982. dren. Child Development, 1982, 53, 920-923.
17. Felleman, E. S., Barden, R. C., Carlson, C. R., Ro- Barnett, M. A., Mathews, K. A., & Howard, J. A. Re-
senberg, L., & Masters, J. C. Children's and adults' lationship between competitiveness and empathy in
recognition of emotional states in young children. 6- and 7-year olds. Developmental Psychology, 1979,
Paper presented at the Annual Meeting of the 15, 221-222.
American Psychological Association, August 1981. Barrcra, M. E., & Maurer, D. The perception of facial
18. Krauss, R. M., & Morency, N. Children's nonverbal expressions by the three-month-old. Child Develop-
encoding and decoding of affect. Paper presented at ment, 1981,52, 203-206.
the Biennial Meeting of the Society for Research in Baumrind, D. Current patterns of parental authority.
Child Development, San Francisco, March 1979. Developmental Psychology Monographs, 1971, 41(1,
19. White, J. B., & Feldman, R. S. Children's decoding Pt 2).
ability and role-taking skills. Paper presented at the Baumrind, D. New directions in socialization research.
Biennial Meeting of the Society for Research in American Psychologist, 1980, 7, 639-652.
Child Development, Boston, April 1981. Herman, P. W. Are women more responsive than men
20. Peraino, J. M., & Sawin, D. B. Empathic distress to the young? A review of developmental and situa-
as a mediator of prosocial behavior. Unpublished tional variables. Psychological Bulletin, 1980,88, 668-
manuscript, University of Texas, Austin, 1982. 695.
21. lannotti, R. J. A longitudinal investigation of role Berger, S. M. Conditioning through vicarious instigation.
taking, altruism, and empathy. Paper presented at Psychological Review, 1962, 69, 450-466.
the Biennial Meeting of the Society for Research in Blanck, P. D., Rosenthal, R., Snodgrass, S. E., DePaulo,
Child Development, New Orleans, March 1977. B. M., & Zuckerman, M. Sex differences in eaves-
dropping on nonverbal cues: Developmental changes.
References Journal of Personality and Social Psychology, 1981,
41, 391-396.
Adams, G. R., Schvaneveldt, J. D., & Jenson, G. O. Sex, Block, J. H. Conceptions of sex role: Some crosscultural
age and perceived competency as correlates of em- and longitudinal perspectives. American Psychologist,
pathic ability in adolescence. Adolescence, 1979, 14, 1973, 28. 512-526.
811-818. Block, J. H. Assessing sex differences: Issues, problems,
128 NANCY EISENBERG AND RANDY LENNON

and pitfalls. Merrill-Palmer Quarterly, 1976, 22, 283- Dymond, R. F. A scale for the measurement of empathic
308. ability. Journal of Consulting Psychology, 1949, 13,
Borke, H. Interpersonal perception of young children: 27-33.
Egocentrism or empathy. Developmental Psychology, Dymond, R. F. Personality and empathy. Journal of Con-
1971,5. 262-269. sulting Psychology, 1950, 14, 343-350.
Borke, H. The development of empathy in Chinese and Eisenberg, N. The development of prosocial moral judg-
American children between three and six years of age. ment and its correlates. (Doctoral dissertation, Uni-
Developmental Psychology, 1973, 9, 102-108. versity of California, Berkeley, 1976). Dissertation
Brehm, S. S., Powell, L. K., & Coke, J. S. The effects of Abstracts International, 1977, 37, 4753B. (University
empathic instructions upon donating behavior: Sex Microfilms No. 77-444).
differences in young children. Sex Roles, in press. Eisenberg, N., Lennon, R., & Roth, K. Prosocial devel-
Bryant, B. An index of empathy for children and ado- opment: A longitudinal study. Developmental Psy-
lescents. Child Development, 1982, 53, 413-425. chology, in press.
Buck, R. A test of nonverbal receiving ability: Prelimi- Eisenberg-Berg, N. The development of children's pro-
nary studies. Human Development Research, 1976, 2, social moral judgment. Developmental Psychology,
162-171. 1979, 75, 128-137.
Buck, R. Nonverbal communication of affect in pre- Eisenberg-Berg, N., & Hand, M. The relationship of pre-
school children: Relationships with personality and schoolers' reasoning about prosocial moral conflicts
skin conductance. Journal of Personality and Social to prosocial behavior. Child Development, 1979, 50,
Psychology, 1977, 35, 225-236. 356-363.
Buck, R. The evolution and development of emotion Eisenberg-Berg, N., & Lennon, R. Altruism and the as-
expression and communication. In S. S. Brehm, S. M. sessment of empathy in the preschool years. Child
Kassin, & F. X. Gibbons (Eds.), Developmental social Development, 1980, 51, 552-557.
psychology. New York: Oxford University Press, 1981. Eisenberg-Berg, N., & Mussen, P. Empathy and moral
Buck, R., Miller, R. E., & Caul, W. F. Sex, personality, development in adolescence. Developmental Psychol-
and physiological variables in the communication of ogy, 1978, 14, 185-186.
affect via facial expression. Journal of Personality and Eisenberg-Berg, N., & Neal, C. Children's moral reason-
Social Psychology, 1974, 30, 587-596. ing about their spontaneous prosocial behavior. De-
Buck, R. W., Savin, V. J., Miller, R. E., & Caul, W. F. velopmental Psychology, 1979, 15, 228-229.
Communication of affect through facial expression in Eisenberg-Berg, N., & Neal, C. The effects of person of
humans. Journal of Personality and Social Psychology, the protagonist and costs of helping on children's
1972, 23, 362-371. moral judgment. Personality and Social Psychology
Cohen, J. Statistical power analyses for the behavioral Bulletin, 1981, 7, 17-23.
sciences (2nd ed.). New York: Academic Press, 1977. Eisenberg-Berg, N., & Roth, K. The development of chil-
Coke, J. S., Batson, C. D., & McDavis, K. Empathic dren's prosocial moral judgment: A longitudinal fol-
mediation of helping: A two-stage model. Journal of low-up. Developmental Psychology, 1980, 16, 375-
Personality and Social Psychology, 1978,56, 752-766. 376.
Cooley, C. H. Two major works: Social organization and Ekman, P., & Friesen, W. V. Constants across culture in
human nature and the social order. Glencoe, 111: Free the face and emotion. Journal of Personality and So-
Press, 1956. (Originally published, 1902) cial Psychology Bulletin, 1971, 17, 124-129.
Cooper, H. M. Statistically combining independent stud- Erikson, E. H. Inner and outer space: Reflections of
ies: A meta-analysis of sex differences in conformity womanhood. In R. J. Lifton (Ed.), The woman in
research. Journal of Personality and Social Psychology, America. Boston: Houghton Mifflin, 1965.
1979,57, 131-146.
Fay, B. The relationships of cognitive moral judgment,
Craig, K. D., & Lowery, H. J. Heart-rate components generosity, and empathic behavior in six- and eight-
of conditioned vicarious autonomic responses. Jour-
year-old children. (Doctoral Dissertation, University
nal of Personality and Social Psychology, 1969, //,
of California, Los Angeles, 1970). Dissertation Ab-
381-387.
stracts International, 1971, 31, 3951A. (University
DePaulo, B. M., Jordan, A., Irvine, A., & Laser, P. S.
Microfilms No. 71-4868,122).
Age changes in the detection of deception. Child De-
velopment, 1982, 53, 701-709. Ferguson, G. A. Statistical analyses in psychology and
Deutsch, F. Effects of sex of subject and story character education (2nd ed.) New York: McGraw Hill, 1966.
on preschoolers' perceptions of affective responses and Feshbach, N. D. Studies of empathic behavior in chil-
interpersonal behavior in story sequences. Develop- dren. In B. A. Maher (Ed.), Progress in experimental
mental Psychology, 1975,77, 112-113. personality research (Vol. 8). New York: Academic
Deutsch, F., & Madle, R. A. Empathy: Historic and cur- Press, 1978.
rent conceptualizations, measurement, and a cogni- Feshbach, N. D. Sex differences in empathy and social
tive theoretical perspective. Human Development, behavior in children. In N. Eisenberg (Ed.), The de-
1975, 18, 267-287. velopment of prosocial behavior. New York: Academic
Deutsch, H. The psychology of women: A psychoanalytic Press, 1982.
interpretation. New York: Grune and Stratton, 1944. Feshbach, N., & Feshbach, S. The relation between em-
Dickey, E. C., & Knower, F. H. A note on some etho- pathy and aggression in two age groups. Developmen-
logical differences in recognition of simulated expres- tal Psychology, 1969, 1, 102-107.
sions of the emotions. American Journal of Sociology, Feshbach, N. D., & Roe, K. Empathy in six and seven
1941, 47, 190-193. year olds. Child Development, 1968,59, 133-145.
SEX DIFFERENCES IN EMPATHY 129

Ford, M. E. The construct validity of egocentrism. Psy- (Ed.), Measuring emotions in infants and children.
chological Bulletin. 1979,56, 1169-1188. New York: Cambridge University Press, 1982.
Foushee, H. C., Davis, M. H., & Archer, R. L. Empathy, Hoffman, M. L., & Levine, L. E. Early sex differences
masculinity, and femininity. JSAS: Catalogue of Se- in empathy. Developmental Psychology, 1976, 6, 557-
lected Documents in Psychology, 1979, 9, 85. (MS# 558.
1974) Hogan, R. Development of an empathy scale. Journal
Freud, S. Some psychical consequences of the anatomical of Consulting and Clinical Psychology, 1969, 33, 307-
distinction between the sexes. In J. Strachey (Ed. & 316.
trans.), Standard edition of the complete psychological Hudson, L. M. On the coherence of role-taking abilities:
works ofSigmund Freud (Vol. 19). London: Hogarth An alternative to correlational analyses. Child Devel-
Press, 1961. (Originally published, 1925.) opment, 1978, 49, 223-227.
Friedman, S. L., & Jacobs, B. S. Sex differences in ne- Hughes, R., Jr., Tingle, B. A., & Sawin, D. B. Devel-
onates' behavioral responsiveness to repeated auditory opment of empathic understanding in children. Child
stimuli. Infant Behavior and Development, 1981, 4, Development, 1981,52, 122-128.
175-183. Hume, D. An inquiry concerning the principles of morals.
Frodi, A. M., & Lamb, M. E. Sex differences in respon- New York: Liberal Arts Press, 1957. (Originally pub-
siveness to infants: A developmental study of psycho- lished, 1751.)
physiological and behavioral responses. Child Devel- lannotti, R. J. Effect of role-taking experiences on role
opment, 1978, 19, 1182-1188. taking, empathy, altruism, and aggression. Develop-
Frodi, A. M., Lamb, M. E., Leavitt, L. A., & Donovan, mental Psychology, 1978, 14, 119-124.
W. L. Fathers' and mothers' responses to infant smiles Isaacs, S. S. Social development in young children. Lon-
and cries, Infant Behavior and Development, 1978, 1, don: Routledge, 1933.
187-198. Kalliopuska, M. Children's helping behaviour: Person-
Frodi, A. M., Lamb, M. E., Leavitt, L. A., Donovan, ality factors and parental influences related to helping
W. L., Neff, C., & Sherry, D. Fathers' and mothers' behaviour. Dissertationes Humanarum Litterarum,
responses to the faces and cries of normal and pre- 24. Helsinki: Academia Scientiarum Fennica, 1980.
mature infants. Developmental Psychology, 1978, 14, Katz, R. L. Empathy: Its nature and uses. London: Free
490-498. Pi-ess, 1963.
Gates, G. S. An experimental study of the growth of Kellogg, W. N., & Eagleson, B. M. The growth of social
social perception. Journal of Educational Psychology, perception in different racial groups. Journal of Ed-
1923,74449-4611. ucational Psychology, 1931, 22, 367-375.
Gitter, A, G., Mostofsky, D. I., & Quincy, A. J., Jr. Race Krebs, D. Empathy and altruism. Journal of Personality
and sex differences in the child's perception of emo- and Social Psychology, 1915,32, 1134-1146.
tion. Child Development, 1971, 42, 2071-2075. Kurdek, L. A., & Rodgon, M. M. Perceptual, cognitive,
Glass, G. V. Primary, secondary, and rneta-analysis of and affective perspective taking in kindergarten through
research. Educational Researcher, 1976, 5, 3-8. sixth-grade children. Developmental Psychology, 1975,
Glass, G, V., McGaw, B,, & Smith, M. L. -Meta-analyses 11, 643-650.
in social research. Beverly Hills: Sage, 1981. Kurtz, C. A., & Eisenberg, N. Role-taking, empathy, and
Hall, J. A. Gender effects in decoding nonverbal cues. resistance to deviation in children. Journal of Genetic
Psychological Bulletin, 1978, 85, 845-858. Psychology, in press.
Hamilton, M. L. Imitative behavior and expressive abil- LaBarbera, J. D., Izard, C. E., Vietze, P., & Parisi, S. A.
ity in facial expression of emotion. Developmental Four- and six-month-old infants' visual responses to
Psychology, 1973, 8, 138. joy, anger, and neutral expressions. Child Develop-
Hebda, M. E., Peterson, R. A., & Miller, L. K. Aggression ment, 1976, 47, 535-538.
anxiety, perception of aggressive cues, and expected Lang, P. J. The application of psychophysiological meth-
retaliation. Developmental Psychology, 1972, 7, 85. ods to the study of psychotherapy and behavior mod-
Hedges, L. V. Estimation of effect size from a series of ification. In A. E. Bergin & S. L. Garfield (Eds.), Hand-
independent experiments. Psychological Bulletin, 1982, book of psychotherapy and behavior change: An em-
92, 490-499. pirical analysis. New York: Wiley, 1971.
Hoffman, M. L. Empathy, role-taking, guilt, and devel- Levine, L. E., & Hoffman, M. L. Empathy and coop-
opment of altruistic motives. In T. Lickona (Ed.), eration in 4-year-olds. Developmental Psychology, 1975,
Moral development and behavior: Theory, research and 4, 533-534.
social issues. New York: Holt, Rinehart & Winston, Liebhart, E. H. Empathy and emergency helping: The
1976. effects of personality, self-concern, and acquaintance.
Hoffman, M. L. Sex differences in empathy and related Journal of Experimental Social Psychology, 1972, 8,
behaviors. Psychological Bulletin, 1977, 54, 712-722. 404-411.
Hoffman, M. L. Is altruism part of human nature? Jour- Lipsitt, L. P., & Levy, N. Electrotactual threshold in the
nal of Personality and Social Psychology, 1981, 40, neonate. Child Development, 1959, 30, 547-554.
121-137. Maccoby, E. E., & Jacklin, C. N. The psychology of sex
Hoffman, M. L. Development of prosocial motivation: differences. Stanford: Stanford University Press, 1974.
Empathy and guilt. In N. Eisenberg (Ed.), The devel- Marcus, R. F, Telleen, S., & Roke, E. J. Relation between
opment of prosocial behavior. New York: Academic cooperation and empathy in young children. Devel-
Press, 1982. opmental Psychology, 1979, 15, 346-347.
Hoffman, M. L. Measurement of empathy. In C. Izard Marks, E. L., Penner, L. A., & Stone, A. V. W. Helping
130 NANCY EISENBERG AND RANDY LENNON

as a function of empathic responses and sociopathy. Staub, D. Bar-Tal, J. Karylowski, & J. Reykowski
Journal of Research in Personality, 1982, 16, 1-20. (Eds.), Development and maintenance ofprosocial be-
Marsh, D. T., Serafica, F., & Barenboim, C. Effect of havior: International perspectives. Plenum Press, in
perspective-taking training on interpersonal problem press.
solving. Child Development, 1980, 57, 140-145. Roe, K. V. A study of empathy in young Greek and U.S.
Marsh, D. T., Serafica, F., & Barenboim, C. Interrela- children. Journal of Cross-Cultural Psychology, 1977,
tionships among perspective taking, interpersonal 8, 493-502.
problem solving, and interpersonal functioning. Jour- Rosenthal, R. Combining results of independent studies.
nal of Genetic Psychology, 1981, 138, 37-48. Psychological Bulletin, 1978, 85, 185-193.
Martin, G. B., & Clark, R. D. HI. Distress crying in Rosenthal, R. The "file drawer problem" and tolerance
neonates: Species and peer specificity. Developmental for null results. Psychological Bulletin, 1979,86, 638-
Psychology, 1982, 18, 3-9. 641.
McDougall, W. An introduction to social psychology Rosenthal, R. On telling tails when combining results
(30th ed.). London: Methuen, 1950. (Originally pub- of independent studies. Psychological Bulletin, 1980,
lished, 1908.) 88, 496-497.
McDougall, W. Outline of psychology. New York: Scrib- Rosenthal, R., & DePaulo, B. M. Sex differences in eaves-
ner, 1923. dropping on nonverbal cues. Journal of Personality
Mead, G. H. Mind, self, and society. Chicago: University and Social Psychology, 1979, 37, 273-285.
of Chicago Press, 1934. Rosenthal, R., Hall, J. A., DiMatteo, M. R., Rogers,
Mehrabian, A. Individual differences in stimulus screen- P. L., & Archer, D. Sensitivity to nonverbal commu-
ing and arousability. Journal of Personality, 1977, 45, nication: The PONS test. Baltimore, Md.: Johns Hop-
237-250. kins University, 1979.
Mehrabian, A., & Epstein, N. A. A measure of emotional Rothenberg, B. Children's social sensitivity and the re-
empathy. Journal of Personality, 1972, 40, 523-543. lationship to interpersonal competence, intrapersonal
Mehrabian, A., & Ksionsky, S. A theory of affiliation. comfort, and intellectual level. Developmental Psy-
Lexington, Mass.: Heath, 1974. chology, 1970, 2, 335-350.
Mood, D. W., Johnson, J. E., & Shantz, C. U. Social Rousseau, J. J. Emile: Or concerning education. Boston:
comprehension and affect-matching in young chil- Heath, 1904. (Originally published, 1762.)
dren. Merrill-Palmer Quarterly, 1978, 24, 63-66. Rubin, K. H. Role taking in childhood: Some method-
Mosteller, F. M., & Bush, R. R. Selected quantitative ological considerations. Child Development, 1978,49,
techniques. In G. Lindzey (Ed.), Handbook of social 428-433.
psychology: Volume 1. Theory and method. Cam- Saarni, C. Children's understanding of display rules for
bridge, Mass.: Addison-Wesley, 1954. expressive behavior. Developmental Psychology, 1979,
Murphy, L. B. Social behavior and child personality. New 75, 424-429.
York: Columbia University, 1937. Sagi, A., & Hoffman, M. L. Empathic distress in new-
Murray, A. D. Infant crying as an activator of altruistic borns. Developmental Psychology, 1976,72, 175-176.
and egoistic motives in adults. Unpublished doctoral Shantz, C. U. The development of social cognition. In
dissertation, Macquarie University, North Ryde, New E. M. Hetherington (Ed.), Review of child development
South Wales, Australia, 1978. research (Vol. 5). Chicago: University of Chicago
Murray, A. D. Aversiveness is in the mind of the beholder: Press, 1975.
Perceptions of infant crying by adults. In B. M. Lester Simner, M. L. Newborn's response to the cry of another
& C. F. Z. Boukydis (Eds.), Infant crying: Theoretical Mam. Developmental Psychology, 1971,5, 136-150.
and research perspectives. New York: Plenum Press, Singer, J. E., Westphal, M., & Niswander, K. R. Sex dif-
in press. ferences in the incidence of neonatal abnormalities
Mussen, P., & Eisenberg-Berg, N. Roots of caring, shar- and abnormal performance in early childhood. Child
ing, and helping: The development ofprosocial behav- Development, 1968, 39, 103-112.
ior in children. San Francisco: Freeman, 1977. Smith, A. The theory of moral sentiment. In H. Schnei-
O'Brien, T. C., & Shapiro, B. J. Statistical significance— der (Ed.), Adam Smith's moral and political philoso-
What? Mathematics Teacher, 1968, 61, 673-676. phy. New York: Hafner 1948. (Originally published,
Odom, R. D., & Lemond, C. M. Developmental differ- 1759.)
ences in the perception and production of facial Staffieri, J. R., feBassett, J. E. Birth order and perception
expressions. Child Development, 1972, 43, 359-369. of facial expressions. Perceptual and Motor Skills,
Olden, C. Notes on the development of empathy. The 1910,30, 606.
Psychoanalytic Study of the Child, 1958, 13, 505-518. Staub, E. Positive social behavior and morality: Social
Oster, H. "Recognition" of emotional expression in in- and personal influences (Vol. 1). New York: Academic
fancy? In M. E. Lamb & L. R. Sherrod (Eds.), Infant Press, 1978.
social cognition. Hillsdale, N.J.: Erlbaum, 1981. Staub, E. Positive social behavior and morality: Social-
Parsons, T., & Bales, R. F. (Eds.), Family, socialization, ization and development (Vol. 2). New York: Academic
and interaction processes. New York: Free Press, 1955. Press, 1979.
Radin, N., & Sagi, A. Childrearing fathers in intact fam- Stern, W. Psychology of early childhood up to the sixth
ilies; II: Israel and the USA. Merrill-Palmer Quarterly, year of age. New York: Holt, 1924.
1982, 28, 111-136. Stotland, E. Exploratory studies of empathy. In L. Ber-
Radke-Yarrow, M., & Zahn-Waxler, C. Roots, motives, kowitz (Ed.), Advances in experimental social psy-
and patterns in children's prosocial behavior. In E. chology (Vol. 4). New York: Academic Press, 1969.
SEX DIFFERENCES IN EMPATHY 131

Stotland, E., & Dunn, R. Empathy, self-esteem, and birth Young-Browne, G., Rosenfeld, H. M., & Horowitz, F. D.
order. Journal of Abnormal and Social Psychology, Infant discrimination of facial expressions. Child De-
1963, 66, 532-544. velopment, 1977, 48, 555-562.
Stotland, E., Mathews, K. E., Sherman, S. E., Hansson, Zeskind, P. S., & Lester, B. M. Acoustic features and
R. O., & Richardson, B. Z. Empathy, fantasy and help- auditory perceptions of the cries of newborns with
ing. Beverly Hills: Sage Publications, 1978. prenatal and perinatal complications. Child Devel-
Stotland, E., & Walsh, J. Birth order in an experimental opment, 1978, 49, 580-589.
study of empathy. Journal of Abnormal and Social Zuckerman, M., Blanck, P. D., DePaulo, B. M., & Ro-
Psychology, 1963, 66, 610-614. senthal, R. Developmental changes in decoding dis-
Stouffer, S. A., Suchman, E. A., DeVinney, L. C., Star, crepant and nondiscrepant nonverbal cues. Develop-
S. A., & Williams, R. M., Jr. The American soldier: mental Psychology, 1980, 16, 220-228.
Adjustment during army life (Vol. 1). Princeton, N.J.: Zuckerman, M., DePaulo, B. M., & Rosenthal, R. Verbal
Princeton University Press, 1949. and nonverbal communication of deception. In L.
Underwood, B., & Moore, B. S. Perspective-taking and Berkowitz (Ed.), Advances in experimental social psy-
altruism. Psychological Bulletin, 1982,97, 143-173. chology (Vol. 14). New York: Academic Press, 1981.
Urberg, K. A., & Docherty, E. M. Development of role- Zuckerman, M., KJorman, R., Larrance, D. T., & Spie-
taking skills in young children. Developmental Psy- gel, N. H. Facial, autonomic, and subjective compo-
chology, 1976, 12, 198-203. nents of emotion: The facial feedback hypotheses ver-
Walton, W. E. Empathic responses in children. Psycho- sus the externalizer-internalizer distinction. Journal of
logical Monographs, 1936, 48(1, Whole No. 213). Personality and Social Psychology, 1981,47, 929-944.
Wilcox, B. M., & Clayton, F. L. Infant visual fixation on Zuckerman, M., & Przewuzman, S. Decoding and en-
motion pictures of the human face. Journal of Ex- coding facial expressions in preschool age children.
perimental Child Psychology, 1968, 6, 22-32. Environmental Psychology and Nonverbal Behavior,
Wispe, L. Sympathy and empathy. In D. L. Sills (Ed.), 1979, 3, 147-163.
International encyclopedia of the social sciences (Vol.
15). New York: Macmillian, 1968.
Wispe, L., Kiecolt, J., & Long, R. E. Demand charac-
teristics, moods, and helping. Social Behavior and Received September 13, 1982
Personality, 1977, 5, 249-255. Revision received January 18, 1983

View publication stats